writing purpose of research paper

Purdue Online Writing Lab Purdue OWL® College of Liberal Arts

Writing a Research Paper

Welcome to the Purdue OWL

This page is brought to you by the OWL at Purdue University. When printing this page, you must include the entire legal notice.

Copyright ©1995-2018 by The Writing Lab & The OWL at Purdue and Purdue University. All rights reserved. This material may not be published, reproduced, broadcast, rewritten, or redistributed without permission. Use of this site constitutes acceptance of our terms and conditions of fair use.

The pages in this section provide detailed information about how to write research papers including discussing research papers as a genre, choosing topics, and finding sources.

The Research Paper

There will come a time in most students' careers when they are assigned a research paper. Such an assignment often creates a great deal of unneeded anxiety in the student, which may result in procrastination and a feeling of confusion and inadequacy. This anxiety frequently stems from the fact that many students are unfamiliar and inexperienced with this genre of writing. Never fear—inexperience and unfamiliarity are situations you can change through practice! Writing a research paper is an essential aspect of academics and should not be avoided on account of one's anxiety. In fact, the process of writing a research paper can be one of the more rewarding experiences one may encounter in academics. What is more, many students will continue to do research throughout their careers, which is one of the reasons this topic is so important.

Becoming an experienced researcher and writer in any field or discipline takes a great deal of practice. There are few individuals for whom this process comes naturally. Remember, even the most seasoned academic veterans have had to learn how to write a research paper at some point in their career. Therefore, with diligence, organization, practice, a willingness to learn (and to make mistakes!), and, perhaps most important of all, patience, students will find that they can achieve great things through their research and writing.

The pages in this section cover the following topic areas related to the process of writing a research paper:

• Genre - This section will provide an overview for understanding the difference between an analytical and argumentative research paper.
• Choosing a Topic - This section will guide the student through the process of choosing topics, whether the topic be one that is assigned or one that the student chooses themselves.
• Identifying an Audience - This section will help the student understand the often times confusing topic of audience by offering some basic guidelines for the process.
• Where Do I Begin - This section concludes the handout by offering several links to resources at Purdue, and also provides an overview of the final stages of writing a research paper.

Writing a Research Paper

This page lists some of the stages involved in writing a library-based research paper.

Although this list suggests that there is a simple, linear process to writing such a paper, the actual process of writing a research paper is often a messy and recursive one, so please use this outline as a flexible guide.

Discovering, Narrowing, and Focusing a Researchable Topic

• Try to find a topic that truly interests you
• Try writing your way to a topic
• Talk with your course instructor and classmates about your topic
• Pose your topic as a question to be answered or a problem to be solved

Finding, Selecting, and Reading Sources

You will need to look at the following types of sources:

• library catalog, periodical indexes, bibliographies, suggestions from your instructor
• primary vs. secondary sources
• journals, books, other documents

Grouping, Sequencing, and Documenting Information

The following systems will help keep you organized:

• a system for noting sources on bibliography cards
• a system for organizing material according to its relative importance
• a system for taking notes

Writing an Outline and a Prospectus for Yourself

Consider the following questions:

• What is the topic?
• Why is it significant?
• What background material is relevant?
• What is my thesis or purpose statement?
• What organizational plan will best support my purpose?

Writing the Introduction

In the introduction you will need to do the following things:

• present relevant background or contextual material
• define terms or concepts when necessary
• explain the focus of the paper and your specific purpose
• reveal your plan of organization

Writing the Body

• Use your outline and prospectus as flexible guides
• Build your essay around points you want to make (i.e., don’t let your sources organize your paper)
• Integrate your sources into your discussion
• Summarize, analyze, explain, and evaluate published work rather than merely reporting it
• Move up and down the “ladder of abstraction” from generalization to varying levels of detail back to generalization

Writing the Conclusion

• If the argument or point of your paper is complex, you may need to summarize the argument for your reader.
• If prior to your conclusion you have not yet explained the significance of your findings or if you are proceeding inductively, use the end of your paper to add your points up, to explain their significance.
• Move from a detailed to a general level of consideration that returns the topic to the context provided by the introduction.
• Perhaps suggest what about this topic needs further research.

Revising the Final Draft

• Check overall organization : logical flow of introduction, coherence and depth of discussion in body, effectiveness of conclusion.
• Paragraph level concerns : topic sentences, sequence of ideas within paragraphs, use of details to support generalizations, summary sentences where necessary, use of transitions within and between paragraphs.
• Sentence level concerns: sentence structure, word choices, punctuation, spelling.
• Documentation: consistent use of one system, citation of all material not considered common knowledge, appropriate use of endnotes or footnotes, accuracy of list of works cited.

Academic and Professional Writing

This is an accordion element with a series of buttons that open and close related content panels.

Analysis Papers

Reading Poetry

A Short Guide to Close Reading for Literary Analysis

Using Literary Quotations

Play Reviews

Writing a Rhetorical Précis to Analyze Nonfiction Texts

Incorporating Interview Data

Grant Proposals

Planning and Writing a Grant Proposal: The Basics

Additional Resources for Grants and Proposal Writing

Job Materials and Application Essays

Writing Personal Statements for Ph.D. Programs

• Before you begin: useful tips for writing your essay
• Guided brainstorming exercises
• Get more help with your essay
• Frequently Asked Questions

Resume Writing Tips

CV Writing Tips

Cover Letters

Business Letters

Proposals and Dissertations

Resources for Proposal Writers

Resources for Dissertators

Research Papers

Planning and Writing Research Papers

Quoting and Paraphrasing

Writing Annotated Bibliographies

Creating Poster Presentations

Writing an Abstract for Your Research Paper

Thank-You Notes

Advice for Students Writing Thank-You Notes to Donors

Reading for a Review

Critical Reviews

Writing a Review of Literature

Scientific Reports

Scientific Report Format

Sample Lab Assignment

Writing for the Web

Writing an Effective Blog Post

Writing for Social Media: A Guide for Academics

What Is a Research Paper?

• An Introduction to Punctuation

Olivia Valdes was the Associate Editorial Director for ThoughtCo. She worked with Dotdash Meredith from 2017 to 2021.

• B.A., American Studies, Yale University

A research paper is a common form of academic writing . Research papers require students and academics to locate information about a topic (that is, to conduct research ), take a stand on that topic, and provide support (or evidence) for that position in an organized report.

The term research paper may also refer to a scholarly article that contains the results of original research or an evaluation of research conducted by others. Most scholarly articles must undergo a process of peer review before they can be accepted for publication in an academic journal.

Define Your Research Question

The first step in writing a research paper is defining your research question . Has your instructor assigned a specific topic? If so, great—you've got this step covered. If not, review the guidelines of the assignment. Your instructor has likely provided several general subjects for your consideration. Your research paper should focus on a specific angle on one of these subjects. Spend some time mulling over your options before deciding which one you'd like to explore more deeply.

Try to choose a research question that interests you. The research process is time-consuming, and you'll be significantly more motivated if you have a genuine desire to learn more about the topic. You should also consider whether you have access to all of the resources necessary to conduct thorough research on your topic, such as primary and secondary sources .

Create a Research Strategy 

Approach the research process systematically by creating a research strategy. First, review your library's website. What resources are available? Where will you find them? Do any resources require a special process to gain access? Start gathering those resources—especially those that may be difficult to access—as soon as possible.

Second, make an appointment with a reference librarian . A reference librarian is nothing short of a research superhero. He or she will listen to your research question, offer suggestions for how to focus your research, and direct you toward valuable sources that directly relate to your topic.

Evaluate Sources

Now that you've gathered a wide array of sources, it's time to evaluate them. First, consider the reliability of the information. Where is the information coming from? What is the origin of the source? Second, assess the  relevance  of the information. How does this information relate to your research question? Does it support, refute, or add context to your position? How does it relate to the other sources you'll be using in your paper? Once you have determined that your sources are both reliable and relevant, you can proceed confidently to the writing phase. 

Why Write Research Papers? 

The research process is one of the most taxing academic tasks you'll be asked to complete. Luckily, the value of writing a research paper goes beyond that A+ you hope to receive. Here are just some of the benefits of research papers. 

• Learning Scholarly Conventions:  Writing a research paper is a crash course in the stylistic conventions of scholarly writing. During the research and writing process, you'll learn how to document your research, cite sources appropriately, format an academic paper, maintain an academic tone, and more.
• Organizing Information: In a way, research is nothing more than a massive organizational project. The information available to you is near-infinite, and it's your job to review that information, narrow it down, categorize it, and present it in a clear, relevant format. This process requires attention to detail and major brainpower.
• Managing Time: Research papers put your time management  skills to the test. Every step of the research and writing process takes time, and it's up to you to set aside the time you'll need to complete each step of the task. Maximize your efficiency by creating a research schedule and inserting blocks of "research time" into your calendar as soon as you receive the assignment. 
• Exploring Your Chosen Subject:  We couldn't forget the best part of research papers—learning about something that truly excites you. No matter what topic you choose, you're bound to come away from the research process with new ideas and countless nuggets of fascinating information. 

The best research papers are the result of genuine interest and a thorough research process. With these ideas in mind, go forth and research. Welcome to the scholarly conversation!

• How to Narrow the Research Topic for Your Paper
• 10 Places to Research Your Paper
• What Is a Senior Thesis?
• How to Write a Research Paper That Earns an A
• Documentation in Reports and Research Papers
• What Is a Bibliography?
• 5 Steps to Writing a Position Paper
• How to Organize Research Notes
• How to Develop a Research Paper Timeline
• An Introduction to Academic Writing
• Writing an Annotated Bibliography for a Paper
• Abstract Writing for Sociology
• Writing a Paper about an Environmental Issue
• Finding Trustworthy Sources
• Topic In Composition and Speech
• How to Write a 10-Page Research Paper

By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts.

Home » Research Paper – Structure, Examples and Writing Guide

Research Paper – Structure, Examples and Writing Guide

Table of Contents

Research Paper

Definition:

Research Paper is a written document that presents the author’s original research, analysis, and interpretation of a specific topic or issue.

It is typically based on Empirical Evidence, and may involve qualitative or quantitative research methods, or a combination of both. The purpose of a research paper is to contribute new knowledge or insights to a particular field of study, and to demonstrate the author’s understanding of the existing literature and theories related to the topic.

Structure of Research Paper

The structure of a research paper typically follows a standard format, consisting of several sections that convey specific information about the research study. The following is a detailed explanation of the structure of a research paper:

The title page contains the title of the paper, the name(s) of the author(s), and the affiliation(s) of the author(s). It also includes the date of submission and possibly, the name of the journal or conference where the paper is to be published.

The abstract is a brief summary of the research paper, typically ranging from 100 to 250 words. It should include the research question, the methods used, the key findings, and the implications of the results. The abstract should be written in a concise and clear manner to allow readers to quickly grasp the essence of the research.

Introduction

The introduction section of a research paper provides background information about the research problem, the research question, and the research objectives. It also outlines the significance of the research, the research gap that it aims to fill, and the approach taken to address the research question. Finally, the introduction section ends with a clear statement of the research hypothesis or research question.

Literature Review

The literature review section of a research paper provides an overview of the existing literature on the topic of study. It includes a critical analysis and synthesis of the literature, highlighting the key concepts, themes, and debates. The literature review should also demonstrate the research gap and how the current study seeks to address it.

The methods section of a research paper describes the research design, the sample selection, the data collection and analysis procedures, and the statistical methods used to analyze the data. This section should provide sufficient detail for other researchers to replicate the study.

The results section presents the findings of the research, using tables, graphs, and figures to illustrate the data. The findings should be presented in a clear and concise manner, with reference to the research question and hypothesis.

The discussion section of a research paper interprets the findings and discusses their implications for the research question, the literature review, and the field of study. It should also address the limitations of the study and suggest future research directions.

The conclusion section summarizes the main findings of the study, restates the research question and hypothesis, and provides a final reflection on the significance of the research.

The references section provides a list of all the sources cited in the paper, following a specific citation style such as APA, MLA or Chicago.

How to Write Research Paper

You can write Research Paper by the following guide:

• Choose a Topic: The first step is to select a topic that interests you and is relevant to your field of study. Brainstorm ideas and narrow down to a research question that is specific and researchable.
• Conduct a Literature Review: The literature review helps you identify the gap in the existing research and provides a basis for your research question. It also helps you to develop a theoretical framework and research hypothesis.
• Develop a Thesis Statement : The thesis statement is the main argument of your research paper. It should be clear, concise and specific to your research question.
• Plan your Research: Develop a research plan that outlines the methods, data sources, and data analysis procedures. This will help you to collect and analyze data effectively.
• Collect and Analyze Data: Collect data using various methods such as surveys, interviews, observations, or experiments. Analyze data using statistical tools or other qualitative methods.
• Organize your Paper : Organize your paper into sections such as Introduction, Literature Review, Methods, Results, Discussion, and Conclusion. Ensure that each section is coherent and follows a logical flow.
• Write your Paper : Start by writing the introduction, followed by the literature review, methods, results, discussion, and conclusion. Ensure that your writing is clear, concise, and follows the required formatting and citation styles.
• Edit and Proofread your Paper: Review your paper for grammar and spelling errors, and ensure that it is well-structured and easy to read. Ask someone else to review your paper to get feedback and suggestions for improvement.
• Cite your Sources: Ensure that you properly cite all sources used in your research paper. This is essential for giving credit to the original authors and avoiding plagiarism.

Research Paper Example

Note : The below example research paper is for illustrative purposes only and is not an actual research paper. Actual research papers may have different structures, contents, and formats depending on the field of study, research question, data collection and analysis methods, and other factors. Students should always consult with their professors or supervisors for specific guidelines and expectations for their research papers.

Research Paper Example sample for Students:

Title: The Impact of Social Media on Mental Health among Young Adults

Abstract: This study aims to investigate the impact of social media use on the mental health of young adults. A literature review was conducted to examine the existing research on the topic. A survey was then administered to 200 university students to collect data on their social media use, mental health status, and perceived impact of social media on their mental health. The results showed that social media use is positively associated with depression, anxiety, and stress. The study also found that social comparison, cyberbullying, and FOMO (Fear of Missing Out) are significant predictors of mental health problems among young adults.

Introduction: Social media has become an integral part of modern life, particularly among young adults. While social media has many benefits, including increased communication and social connectivity, it has also been associated with negative outcomes, such as addiction, cyberbullying, and mental health problems. This study aims to investigate the impact of social media use on the mental health of young adults.

Literature Review: The literature review highlights the existing research on the impact of social media use on mental health. The review shows that social media use is associated with depression, anxiety, stress, and other mental health problems. The review also identifies the factors that contribute to the negative impact of social media, including social comparison, cyberbullying, and FOMO.

Methods : A survey was administered to 200 university students to collect data on their social media use, mental health status, and perceived impact of social media on their mental health. The survey included questions on social media use, mental health status (measured using the DASS-21), and perceived impact of social media on their mental health. Data were analyzed using descriptive statistics and regression analysis.

Results : The results showed that social media use is positively associated with depression, anxiety, and stress. The study also found that social comparison, cyberbullying, and FOMO are significant predictors of mental health problems among young adults.

Discussion : The study’s findings suggest that social media use has a negative impact on the mental health of young adults. The study highlights the need for interventions that address the factors contributing to the negative impact of social media, such as social comparison, cyberbullying, and FOMO.

Conclusion : In conclusion, social media use has a significant impact on the mental health of young adults. The study’s findings underscore the need for interventions that promote healthy social media use and address the negative outcomes associated with social media use. Future research can explore the effectiveness of interventions aimed at reducing the negative impact of social media on mental health. Additionally, longitudinal studies can investigate the long-term effects of social media use on mental health.

Limitations : The study has some limitations, including the use of self-report measures and a cross-sectional design. The use of self-report measures may result in biased responses, and a cross-sectional design limits the ability to establish causality.

Implications: The study’s findings have implications for mental health professionals, educators, and policymakers. Mental health professionals can use the findings to develop interventions that address the negative impact of social media use on mental health. Educators can incorporate social media literacy into their curriculum to promote healthy social media use among young adults. Policymakers can use the findings to develop policies that protect young adults from the negative outcomes associated with social media use.

References :

• Twenge, J. M., & Campbell, W. K. (2019). Associations between screen time and lower psychological well-being among children and adolescents: Evidence from a population-based study. Preventive medicine reports, 15, 100918.
• Primack, B. A., Shensa, A., Escobar-Viera, C. G., Barrett, E. L., Sidani, J. E., Colditz, J. B., … & James, A. E. (2017). Use of multiple social media platforms and symptoms of depression and anxiety: A nationally-representative study among US young adults. Computers in Human Behavior, 69, 1-9.
• Van der Meer, T. G., & Verhoeven, J. W. (2017). Social media and its impact on academic performance of students. Journal of Information Technology Education: Research, 16, 383-398.

Appendix : The survey used in this study is provided below.

Social Media and Mental Health Survey

• How often do you use social media per day?
• Less than 30 minutes
• 30 minutes to 1 hour
• 1 to 2 hours
• 2 to 4 hours
• More than 4 hours
• Which social media platforms do you use?
• Others (Please specify)
• How often do you experience the following on social media?
• Social comparison (comparing yourself to others)
• Cyberbullying
• Fear of Missing Out (FOMO)
• Have you ever experienced any of the following mental health problems in the past month?
• Do you think social media use has a positive or negative impact on your mental health?
• Very positive
• Somewhat positive
• Somewhat negative
• Very negative
• In your opinion, which factors contribute to the negative impact of social media on mental health?
• Social comparison
• In your opinion, what interventions could be effective in reducing the negative impact of social media on mental health?
• Education on healthy social media use
• Counseling for mental health problems caused by social media
• Social media detox programs
• Regulation of social media use

Thank you for your participation!

Applications of Research Paper

Research papers have several applications in various fields, including:

• Advancing knowledge: Research papers contribute to the advancement of knowledge by generating new insights, theories, and findings that can inform future research and practice. They help to answer important questions, clarify existing knowledge, and identify areas that require further investigation.
• Informing policy: Research papers can inform policy decisions by providing evidence-based recommendations for policymakers. They can help to identify gaps in current policies, evaluate the effectiveness of interventions, and inform the development of new policies and regulations.
• Improving practice: Research papers can improve practice by providing evidence-based guidance for professionals in various fields, including medicine, education, business, and psychology. They can inform the development of best practices, guidelines, and standards of care that can improve outcomes for individuals and organizations.
• Educating students : Research papers are often used as teaching tools in universities and colleges to educate students about research methods, data analysis, and academic writing. They help students to develop critical thinking skills, research skills, and communication skills that are essential for success in many careers.
• Fostering collaboration: Research papers can foster collaboration among researchers, practitioners, and policymakers by providing a platform for sharing knowledge and ideas. They can facilitate interdisciplinary collaborations and partnerships that can lead to innovative solutions to complex problems.

When to Write Research Paper

Research papers are typically written when a person has completed a research project or when they have conducted a study and have obtained data or findings that they want to share with the academic or professional community. Research papers are usually written in academic settings, such as universities, but they can also be written in professional settings, such as research organizations, government agencies, or private companies.

Here are some common situations where a person might need to write a research paper:

• For academic purposes: Students in universities and colleges are often required to write research papers as part of their coursework, particularly in the social sciences, natural sciences, and humanities. Writing research papers helps students to develop research skills, critical thinking skills, and academic writing skills.
• For publication: Researchers often write research papers to publish their findings in academic journals or to present their work at academic conferences. Publishing research papers is an important way to disseminate research findings to the academic community and to establish oneself as an expert in a particular field.
• To inform policy or practice : Researchers may write research papers to inform policy decisions or to improve practice in various fields. Research findings can be used to inform the development of policies, guidelines, and best practices that can improve outcomes for individuals and organizations.
• To share new insights or ideas: Researchers may write research papers to share new insights or ideas with the academic or professional community. They may present new theories, propose new research methods, or challenge existing paradigms in their field.

Purpose of Research Paper

The purpose of a research paper is to present the results of a study or investigation in a clear, concise, and structured manner. Research papers are written to communicate new knowledge, ideas, or findings to a specific audience, such as researchers, scholars, practitioners, or policymakers. The primary purposes of a research paper are:

• To contribute to the body of knowledge : Research papers aim to add new knowledge or insights to a particular field or discipline. They do this by reporting the results of empirical studies, reviewing and synthesizing existing literature, proposing new theories, or providing new perspectives on a topic.
• To inform or persuade: Research papers are written to inform or persuade the reader about a particular issue, topic, or phenomenon. They present evidence and arguments to support their claims and seek to persuade the reader of the validity of their findings or recommendations.
• To advance the field: Research papers seek to advance the field or discipline by identifying gaps in knowledge, proposing new research questions or approaches, or challenging existing assumptions or paradigms. They aim to contribute to ongoing debates and discussions within a field and to stimulate further research and inquiry.
• To demonstrate research skills: Research papers demonstrate the author’s research skills, including their ability to design and conduct a study, collect and analyze data, and interpret and communicate findings. They also demonstrate the author’s ability to critically evaluate existing literature, synthesize information from multiple sources, and write in a clear and structured manner.

Characteristics of Research Paper

Research papers have several characteristics that distinguish them from other forms of academic or professional writing. Here are some common characteristics of research papers:

• Evidence-based: Research papers are based on empirical evidence, which is collected through rigorous research methods such as experiments, surveys, observations, or interviews. They rely on objective data and facts to support their claims and conclusions.
• Structured and organized: Research papers have a clear and logical structure, with sections such as introduction, literature review, methods, results, discussion, and conclusion. They are organized in a way that helps the reader to follow the argument and understand the findings.
• Formal and objective: Research papers are written in a formal and objective tone, with an emphasis on clarity, precision, and accuracy. They avoid subjective language or personal opinions and instead rely on objective data and analysis to support their arguments.
• Citations and references: Research papers include citations and references to acknowledge the sources of information and ideas used in the paper. They use a specific citation style, such as APA, MLA, or Chicago, to ensure consistency and accuracy.
• Peer-reviewed: Research papers are often peer-reviewed, which means they are evaluated by other experts in the field before they are published. Peer-review ensures that the research is of high quality, meets ethical standards, and contributes to the advancement of knowledge in the field.
• Objective and unbiased: Research papers strive to be objective and unbiased in their presentation of the findings. They avoid personal biases or preconceptions and instead rely on the data and analysis to draw conclusions.

Advantages of Research Paper

Research papers have many advantages, both for the individual researcher and for the broader academic and professional community. Here are some advantages of research papers:

• Contribution to knowledge: Research papers contribute to the body of knowledge in a particular field or discipline. They add new information, insights, and perspectives to existing literature and help advance the understanding of a particular phenomenon or issue.
• Opportunity for intellectual growth: Research papers provide an opportunity for intellectual growth for the researcher. They require critical thinking, problem-solving, and creativity, which can help develop the researcher’s skills and knowledge.
• Career advancement: Research papers can help advance the researcher’s career by demonstrating their expertise and contributions to the field. They can also lead to new research opportunities, collaborations, and funding.
• Academic recognition: Research papers can lead to academic recognition in the form of awards, grants, or invitations to speak at conferences or events. They can also contribute to the researcher’s reputation and standing in the field.
• Impact on policy and practice: Research papers can have a significant impact on policy and practice. They can inform policy decisions, guide practice, and lead to changes in laws, regulations, or procedures.
• Advancement of society: Research papers can contribute to the advancement of society by addressing important issues, identifying solutions to problems, and promoting social justice and equality.

Limitations of Research Paper

Research papers also have some limitations that should be considered when interpreting their findings or implications. Here are some common limitations of research papers:

• Limited generalizability: Research findings may not be generalizable to other populations, settings, or contexts. Studies often use specific samples or conditions that may not reflect the broader population or real-world situations.
• Potential for bias : Research papers may be biased due to factors such as sample selection, measurement errors, or researcher biases. It is important to evaluate the quality of the research design and methods used to ensure that the findings are valid and reliable.
• Ethical concerns: Research papers may raise ethical concerns, such as the use of vulnerable populations or invasive procedures. Researchers must adhere to ethical guidelines and obtain informed consent from participants to ensure that the research is conducted in a responsible and respectful manner.
• Limitations of methodology: Research papers may be limited by the methodology used to collect and analyze data. For example, certain research methods may not capture the complexity or nuance of a particular phenomenon, or may not be appropriate for certain research questions.
• Publication bias: Research papers may be subject to publication bias, where positive or significant findings are more likely to be published than negative or non-significant findings. This can skew the overall findings of a particular area of research.
• Time and resource constraints: Research papers may be limited by time and resource constraints, which can affect the quality and scope of the research. Researchers may not have access to certain data or resources, or may be unable to conduct long-term studies due to practical limitations.

About the author

Muhammad Hassan

Researcher, Academic Writer, Web developer

You may also like

Scope of the Research – Writing Guide and...

Data Verification – Process, Types and Examples

Research Topics – Ideas and Examples

Institutional Review Board – Application Sample...

Delimitations in Research – Types, Examples and...

Research Contribution – Thesis Guide

Want to create or adapt books like this? Learn more about how Pressbooks supports open publishing practices.

13.1 Formatting a Research Paper

Learning objectives.

• Identify the major components of a research paper written using American Psychological Association (APA) style.
• Apply general APA style and formatting conventions in a research paper.

In this chapter, you will learn how to use APA style , the documentation and formatting style followed by the American Psychological Association, as well as MLA style , from the Modern Language Association. There are a few major formatting styles used in academic texts, including AMA, Chicago, and Turabian:

• AMA (American Medical Association) for medicine, health, and biological sciences
• APA (American Psychological Association) for education, psychology, and the social sciences
• Chicago—a common style used in everyday publications like magazines, newspapers, and books
• MLA (Modern Language Association) for English, literature, arts, and humanities
• Turabian—another common style designed for its universal application across all subjects and disciplines

While all the formatting and citation styles have their own use and applications, in this chapter we focus our attention on the two styles you are most likely to use in your academic studies: APA and MLA.

If you find that the rules of proper source documentation are difficult to keep straight, you are not alone. Writing a good research paper is, in and of itself, a major intellectual challenge. Having to follow detailed citation and formatting guidelines as well may seem like just one more task to add to an already-too-long list of requirements.

Following these guidelines, however, serves several important purposes. First, it signals to your readers that your paper should be taken seriously as a student’s contribution to a given academic or professional field; it is the literary equivalent of wearing a tailored suit to a job interview. Second, it shows that you respect other people’s work enough to give them proper credit for it. Finally, it helps your reader find additional materials if he or she wishes to learn more about your topic.

Furthermore, producing a letter-perfect APA-style paper need not be burdensome. Yes, it requires careful attention to detail. However, you can simplify the process if you keep these broad guidelines in mind:

• Work ahead whenever you can. Chapter 11 “Writing from Research: What Will I Learn?” includes tips for keeping track of your sources early in the research process, which will save time later on.
• Get it right the first time. Apply APA guidelines as you write, so you will not have much to correct during the editing stage. Again, putting in a little extra time early on can save time later.
• Use the resources available to you. In addition to the guidelines provided in this chapter, you may wish to consult the APA website at http://www.apa.org or the Purdue University Online Writing lab at http://owl.english.purdue.edu , which regularly updates its online style guidelines.

General Formatting Guidelines

This chapter provides detailed guidelines for using the citation and formatting conventions developed by the American Psychological Association, or APA. Writers in disciplines as diverse as astrophysics, biology, psychology, and education follow APA style. The major components of a paper written in APA style are listed in the following box.

These are the major components of an APA-style paper:

Body, which includes the following:

• Headings and, if necessary, subheadings to organize the content
• In-text citations of research sources
• References page

All these components must be saved in one document, not as separate documents.

The title page of your paper includes the following information:

• Title of the paper
• Author’s name
• Name of the institution with which the author is affiliated
• Header at the top of the page with the paper title (in capital letters) and the page number (If the title is lengthy, you may use a shortened form of it in the header.)

List the first three elements in the order given in the previous list, centered about one third of the way down from the top of the page. Use the headers and footers tool of your word-processing program to add the header, with the title text at the left and the page number in the upper-right corner. Your title page should look like the following example.

The next page of your paper provides an abstract , or brief summary of your findings. An abstract does not need to be provided in every paper, but an abstract should be used in papers that include a hypothesis. A good abstract is concise—about one hundred fifty to two hundred fifty words—and is written in an objective, impersonal style. Your writing voice will not be as apparent here as in the body of your paper. When writing the abstract, take a just-the-facts approach, and summarize your research question and your findings in a few sentences.

In Chapter 12 “Writing a Research Paper” , you read a paper written by a student named Jorge, who researched the effectiveness of low-carbohydrate diets. Read Jorge’s abstract. Note how it sums up the major ideas in his paper without going into excessive detail.

Write an abstract summarizing your paper. Briefly introduce the topic, state your findings, and sum up what conclusions you can draw from your research. Use the word count feature of your word-processing program to make sure your abstract does not exceed one hundred fifty words.

Depending on your field of study, you may sometimes write research papers that present extensive primary research, such as your own experiment or survey. In your abstract, summarize your research question and your findings, and briefly indicate how your study relates to prior research in the field.

Margins, Pagination, and Headings

APA style requirements also address specific formatting concerns, such as margins, pagination, and heading styles, within the body of the paper. Review the following APA guidelines.

Use these general guidelines to format the paper:

• Set the top, bottom, and side margins of your paper at 1 inch.
• Use double-spaced text throughout your paper.
• Use a standard font, such as Times New Roman or Arial, in a legible size (10- to 12-point).
• Use continuous pagination throughout the paper, including the title page and the references section. Page numbers appear flush right within your header.
• Section headings and subsection headings within the body of your paper use different types of formatting depending on the level of information you are presenting. Additional details from Jorge’s paper are provided.

Begin formatting the final draft of your paper according to APA guidelines. You may work with an existing document or set up a new document if you choose. Include the following:

• Your title page
• The abstract you created in Note 13.8 “Exercise 1”
• Correct headers and page numbers for your title page and abstract

APA style uses section headings to organize information, making it easy for the reader to follow the writer’s train of thought and to know immediately what major topics are covered. Depending on the length and complexity of the paper, its major sections may also be divided into subsections, sub-subsections, and so on. These smaller sections, in turn, use different heading styles to indicate different levels of information. In essence, you are using headings to create a hierarchy of information.

The following heading styles used in APA formatting are listed in order of greatest to least importance:

• Section headings use centered, boldface type. Headings use title case, with important words in the heading capitalized.
• Subsection headings use left-aligned, boldface type. Headings use title case.
• The third level uses left-aligned, indented, boldface type. Headings use a capital letter only for the first word, and they end in a period.
• The fourth level follows the same style used for the previous level, but the headings are boldfaced and italicized.
• The fifth level follows the same style used for the previous level, but the headings are italicized and not boldfaced.

Visually, the hierarchy of information is organized as indicated in Table 13.1 “Section Headings” .

Table 13.1 Section Headings

A college research paper may not use all the heading levels shown in Table 13.1 “Section Headings” , but you are likely to encounter them in academic journal articles that use APA style. For a brief paper, you may find that level 1 headings suffice. Longer or more complex papers may need level 2 headings or other lower-level headings to organize information clearly. Use your outline to craft your major section headings and determine whether any subtopics are substantial enough to require additional levels of headings.

Working with the document you developed in Note 13.11 “Exercise 2” , begin setting up the heading structure of the final draft of your research paper according to APA guidelines. Include your title and at least two to three major section headings, and follow the formatting guidelines provided above. If your major sections should be broken into subsections, add those headings as well. Use your outline to help you.

Because Jorge used only level 1 headings, his Exercise 3 would look like the following:

Citation Guidelines

In-text citations.

Throughout the body of your paper, include a citation whenever you quote or paraphrase material from your research sources. As you learned in Chapter 11 “Writing from Research: What Will I Learn?” , the purpose of citations is twofold: to give credit to others for their ideas and to allow your reader to follow up and learn more about the topic if desired. Your in-text citations provide basic information about your source; each source you cite will have a longer entry in the references section that provides more detailed information.

In-text citations must provide the name of the author or authors and the year the source was published. (When a given source does not list an individual author, you may provide the source title or the name of the organization that published the material instead.) When directly quoting a source, it is also required that you include the page number where the quote appears in your citation.

This information may be included within the sentence or in a parenthetical reference at the end of the sentence, as in these examples.

Epstein (2010) points out that “junk food cannot be considered addictive in the same way that we think of psychoactive drugs as addictive” (p. 137).

Here, the writer names the source author when introducing the quote and provides the publication date in parentheses after the author’s name. The page number appears in parentheses after the closing quotation marks and before the period that ends the sentence.

Addiction researchers caution that “junk food cannot be considered addictive in the same way that we think of psychoactive drugs as addictive” (Epstein, 2010, p. 137).

Here, the writer provides a parenthetical citation at the end of the sentence that includes the author’s name, the year of publication, and the page number separated by commas. Again, the parenthetical citation is placed after the closing quotation marks and before the period at the end of the sentence.

As noted in the book Junk Food, Junk Science (Epstein, 2010, p. 137), “junk food cannot be considered addictive in the same way that we think of psychoactive drugs as addictive.”

Here, the writer chose to mention the source title in the sentence (an optional piece of information to include) and followed the title with a parenthetical citation. Note that the parenthetical citation is placed before the comma that signals the end of the introductory phrase.

David Epstein’s book Junk Food, Junk Science (2010) pointed out that “junk food cannot be considered addictive in the same way that we think of psychoactive drugs as addictive” (p. 137).

Another variation is to introduce the author and the source title in your sentence and include the publication date and page number in parentheses within the sentence or at the end of the sentence. As long as you have included the essential information, you can choose the option that works best for that particular sentence and source.

Citing a book with a single author is usually a straightforward task. Of course, your research may require that you cite many other types of sources, such as books or articles with more than one author or sources with no individual author listed. You may also need to cite sources available in both print and online and nonprint sources, such as websites and personal interviews. Chapter 13 “APA and MLA Documentation and Formatting” , Section 13.2 “Citing and Referencing Techniques” and Section 13.3 “Creating a References Section” provide extensive guidelines for citing a variety of source types.

Writing at Work

APA is just one of several different styles with its own guidelines for documentation, formatting, and language usage. Depending on your field of interest, you may be exposed to additional styles, such as the following:

• MLA style. Determined by the Modern Languages Association and used for papers in literature, languages, and other disciplines in the humanities.
• Chicago style. Outlined in the Chicago Manual of Style and sometimes used for papers in the humanities and the sciences; many professional organizations use this style for publications as well.
• Associated Press (AP) style. Used by professional journalists.

References List

The brief citations included in the body of your paper correspond to the more detailed citations provided at the end of the paper in the references section. In-text citations provide basic information—the author’s name, the publication date, and the page number if necessary—while the references section provides more extensive bibliographical information. Again, this information allows your reader to follow up on the sources you cited and do additional reading about the topic if desired.

The specific format of entries in the list of references varies slightly for different source types, but the entries generally include the following information:

• The name(s) of the author(s) or institution that wrote the source
• The year of publication and, where applicable, the exact date of publication
• The full title of the source
• For books, the city of publication
• For articles or essays, the name of the periodical or book in which the article or essay appears
• For magazine and journal articles, the volume number, issue number, and pages where the article appears
• For sources on the web, the URL where the source is located

The references page is double spaced and lists entries in alphabetical order by the author’s last name. If an entry continues for more than one line, the second line and each subsequent line are indented five spaces. Review the following example. ( Chapter 13 “APA and MLA Documentation and Formatting” , Section 13.3 “Creating a References Section” provides extensive guidelines for formatting reference entries for different types of sources.)

In APA style, book and article titles are formatted in sentence case, not title case. Sentence case means that only the first word is capitalized, along with any proper nouns.

Key Takeaways

• Following proper citation and formatting guidelines helps writers ensure that their work will be taken seriously, give proper credit to other authors for their work, and provide valuable information to readers.
• Working ahead and taking care to cite sources correctly the first time are ways writers can save time during the editing stage of writing a research paper.
• APA papers usually include an abstract that concisely summarizes the paper.
• APA papers use a specific headings structure to provide a clear hierarchy of information.
• In APA papers, in-text citations usually include the name(s) of the author(s) and the year of publication.
• In-text citations correspond to entries in the references section, which provide detailed bibliographical information about a source.

Writing for Success Copyright © 2015 by University of Minnesota is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License , except where otherwise noted.

Microsoft 365 Life Hacks > Writing > How to write an introduction for a research paper

How to write an introduction for a research paper

Beginnings are hard. Beginning a research paper is no exception. Many students—and pros—struggle with how to write an introduction for a research paper.

This short guide will describe the purpose of a research paper introduction and how to create a good one.

What is an introduction for a research paper?

Introductions to research papers do a lot of work.

It may seem obvious, but introductions are always placed at the beginning of a paper. They guide your reader from a general subject area to the narrow topic that your paper covers. They also explain your paper’s:

• Scope: The topic you’ll be covering
• Context: The background of your topic
• Importance: Why your research matters in the context of an industry or the world

Your introduction will cover a lot of ground. However, it will only be half of a page to a few pages long. The length depends on the size of your paper as a whole. In many cases, the introduction will be shorter than all of the other sections of your paper.

Write with Confidence using Editor

Elevate your writing with real-time, intelligent assistance

Why is an introduction vital to a research paper?

The introduction to your research paper isn’t just important. It’s critical.

Your readers don’t know what your research paper is about from the title. That’s where your introduction comes in. A good introduction will:

• Help your reader understand your topic’s background
• Explain why your research paper is worth reading
• Offer a guide for navigating the rest of the piece
• Pique your reader’s interest

Without a clear introduction, your readers will struggle. They may feel confused when they start reading your paper. They might even give up entirely. Your introduction will ground them and prepare them for the in-depth research to come.

What should you include in an introduction for a research paper?

Research paper introductions are always unique. After all, research is original by definition. However, they often contain six essential items. These are:

• An overview of the topic. Start with a general overview of your topic. Narrow the overview until you address your paper’s specific subject. Then, mention questions or concerns you had about the case. Note that you will address them in the publication.
• Prior research. Your introduction is the place to review other conclusions on your topic. Include both older scholars and modern scholars. This background information shows that you are aware of prior research. It also introduces past findings to those who might not have that expertise.
• A rationale for your paper. Explain why your topic needs to be addressed right now. If applicable, connect it to current issues. Additionally, you can show a problem with former theories or reveal a gap in current research. No matter how you do it, a good rationale will interest your readers and demonstrate why they must read the rest of your paper.
• Describe the methodology you used. Recount your processes to make your paper more credible. Lay out your goal and the questions you will address. Reveal how you conducted research and describe how you measured results. Moreover, explain why you made key choices.
• A thesis statement. Your main introduction should end with a thesis statement. This statement summarizes the ideas that will run through your entire research article. It should be straightforward and clear.
• An outline. Introductions often conclude with an outline. Your layout should quickly review what you intend to cover in the following sections. Think of it as a roadmap, guiding your reader to the end of your paper.

These six items are emphasized more or less, depending on your field. For example, a physics research paper might emphasize methodology. An English journal article might highlight the overview.

Three tips for writing your introduction

We don’t just want you to learn how to write an introduction for a research paper. We want you to learn how to make it shine.

There are three things you can do that will make it easier to write a great introduction. You can:

• Write your introduction last. An introduction summarizes all of the things you’ve learned from your research. While it can feel good to get your preface done quickly, you should write the rest of your paper first. Then, you’ll find it easy to create a clear overview.
• Include a strong quotation or story upfront. You want your paper to be full of substance. But that doesn’t mean it should feel boring or flat. Add a relevant quotation or surprising anecdote to the beginning of your introduction. This technique will pique the interest of your reader and leave them wanting more.
• Be concise. Research papers cover complex topics. To help your readers, try to write as clearly as possible. Use concise sentences. Check for confusing grammar or syntax . Read your introduction out loud to catch awkward phrases. Before you finish your paper, be sure to proofread, too. Mistakes can seem unprofessional.

Get started with Microsoft 365

It’s the Office you know, plus the tools to help you work better together, so you can get more done—anytime, anywhere.

Topics in this article

More articles like this one.

Playing with paraprosdokians

There are three types of people in this world: those who can count, and those who can’t. That’s a paraprosdokian right there. Learn how to take a phrase and turn it on its head.

Is it OK to use AI on your school assignments?

Generative AI is a popular and powerful tool, but it’s not without its pitfalls. Learn how to avoid consequences, such as plagiarism, when writing with AI.

6 types of rhyming schemes to use in your work

Depending on how you use them, rhyme has the power to create joy or suspense. Learn about six different types of rhyme to try out next time you sit down to write.

6 different poetry forms to play with

Did you know trying new poetry forms doesn’t just make you a better poet? It makes you a better writer too. Try out one of these six poetry types to see what you can come up with—bonus points if you try all six.

Everything you need to achieve more in less time

Get powerful productivity and security apps with Microsoft 365

Explore Other Categories

In order to continue enjoying our site, we ask that you confirm your identity as a human. Thank you very much for your cooperation.

• Affiliate Program

• UNITED STATES
• 台灣 (TAIWAN)
• TÜRKIYE (TURKEY)
• Academic Editing Services
• - Research Paper
• - Journal Manuscript
• - Dissertation
• - College & University Assignments
• Admissions Editing Services
• - Application Essay
• - Personal Statement
• - Recommendation Letter
• - Cover Letter
• - CV/Resume
• Business Editing Services
• - Business Documents
• - Report & Brochure
• - Website & Blog
• Writer Editing Services
• - Script & Screenplay
• Our Editors
• Client Reviews
• Editing & Proofreading Prices
• Wordvice Points
• Partner Discount
• Plagiarism Checker
• APA Citation Generator
• MLA Citation Generator
• Chicago Citation Generator
• Vancouver Citation Generator
• - APA Style
• - MLA Style
• - Chicago Style
• - Vancouver Style
• Writing & Editing Guide
• Academic Resources
• Admissions Resources

Writing an Effective Research Paper: Structure & Content

Essential Guidelines for Structuring a Research Paper

Lecturer: kevin j. heintz, m.a. english.

This lecture was presented at ChungAng University in Seoul, South Korea in November 2018. Wordvice/Essay Review Managing Editor Kevin J. Heintz explains how to organize and compose a research manuscript that will get your study published in top journals.

Even researchers whose first language is English must learn some specific rules and follow some standard conventions when writing research papers. This takes a completely different skillset than essay writing or sending emails to your professors and friends, and therefore it is a good idea for every researcher to keep learning how to improve research writing.

Research is about more than just the scientific principles and discoveries you are making—it is about sharing these discoveries with fellow researchers and with the public. And to do this, researchers must publish their work in journals. Strong writing is key to making your research more accessible and powerful, and therefore this presentation is not about the rigors of research, but the demands of research writing. The methods and information in this lecture can be applied to almost any kind of research paper, although of course the exact structure and content will be somewhat determined by where you are submitting your research.

Lecture Content

• Overview of Research Paper Writing
• The Structure of a Research Paper
• Composing Your Paper Sections
• Tips for Improving Quality of Writing

*Quizzes are given throughout the lecture to test your comprehension and understanding.

Research Paper Structure Overview

“what should a research paper do”.

• Share the knowledge you have gained about a specific area of study with other researchers
• Show how your study fits into current science.
• Inform the public about important scientific activity.
• Explain clearly and succinctly the context of your study, including relevant literature (Introduction), the methods used for research and analysis (Methods), the findings of your study (Results), and the implications for these results and further research that might be needed (Discussion and Conclusion).

“What are the most important factors to consider when writing a research paper?”

The research you conduct should of course be novel, timely, rigorous, and hopefully interesting. But you must also transmit your scientific research into  writing —a well-written paper will greatly improve your chances of getting accepted into journals. Here is an overview of the factors that help create quality writing in a research paper:

• All of the parts of your paper should fit together in an order that makes sense.
• Include all necessary information in each section needed to understand the other sections.
• Do not repeat information unless it is necessary.
• Ensure that your sentences are grammatically and logically coherent.

Organization

• Most scientific papers follow the  IMRD  structure—be sure to put the right parts in the right section (e.g., don’t include the literature review in the Methods section).
• As you do research you will notice that there are a great many pieces of information and data you COULD include in your paper. However, you need to conform to length guidelines and keep your paper focused. Therefore, you should be sure that you are choosing a proper number of items to focus on for each section.
• For example, if your study has 10 results but your paper can only be 4,000 words, you might want to narrow down these results to only those that support your hypothesis, perhaps the 3-5 most important results.
• The same applies to the Introduction, where you must choose what background, context, and relevant literature to include. Be sure to only include information that gives readers a focused and relevant understanding of your area of study.
• Clarity is related to coherence, organization, and relevance. It means ensuring that each paragraph and sentence in your paper is natural and easy to read and understand: proper grammar, phrasing, and style are key to writing a paper that is readable and comprehensible to both experts and possibly non-experts, depending on your target audience.
• Perhaps the most important rule is to  conform to the formatting guidelines and other style conventions of the journal to which you are submitting.  Check the “GUIDE FOR AUTHORS” section of the journal or conference, or if the paper is for a class, ensure that you are using the proper formatting requirements. Here is one handy site:  OWL—Online Writing Lab at Purdue University

Research Paper Structure

The general structure of scientific research papers is IMR&D (Introduction, Methods, Results, and Discussion). The information moves from broad to specific to broad again as seen in this diagram, the Introduction and Discussion taking up the most room in your paper and the Methods and Results usually being the shortest ad most focused sections. However, the order in which you write your paper will not be the same as the final order of the information. Let’s first look briefly at what each section does and then discuss how to organize and compose your work.

Introduction Section

What does it do.

*Discusses the problem to be solved (purpose statement)

*Describes where your research fits into the current science (background and context)

*Uses primary literature with citations and summarizes the current understanding of the problem (“literature review”)

When do you write it?

*Write it last—after the conclusion and before the title and abstract

Methods Section

*Tells how you did the study—what materials and methods of research and analysis were used.

*First section you write—after preparing your figures and tables

Results Section

What does it do.

*Explains the important findings of your study that help to answer your research question or hypothesis and address your purpose statement.

*After the Methods and before the Discussion/Conclusion

Discussion/Conclusion Section

*Explains what your findings mean and what the implications and importance are both to your specific area of research and in a broader context (i.e., to the wider field or to society ).

*Includes limitations to your study and discusses possible future research that is needed to answer your research question more clearly and address closely related questions.

*After the Results Section and before the Introduction

Composing Your Research Paper Sections

This portion of the lecture focuses on developing techniques for composing your paper. You should always go back through your paper after one section is finished and correct or change another part, but by composing in this order you will be sure to include all of the important information. Not that the Methods and Results sections are written first. The reason for this is because you will not be changing or adding to these sections after you have evaluated your research—they represent the core data of your study.

Step 1: Prepare the figures and tables

Most likely, your research paper will use some figures, tables, or other graphics—they are also core data because they are usually numbers representing your findings and methods used. We won’t go into the details of how to prepare these here, but in the  Results section , we will go over how to write captions for the figures based on the data and research questions. For a detailed explanation of preparing and formatting figures, check out these sites (every journal will have their own formatting guidelines):

• Springer Online Research Resources
• ACSESS Digital Library (ASA, CSSA, and SSSA publications for reference)

Step 2: Write the Method s section

This section responds to the question  “How was the problem studied and analyzed?”

The Methods section should:

• Describe how an experiment was done
• Give a rationale for why specific experimental procedures were chosen
• Describe what was done to answer the research question and how it was done.
• Explain how the results were analyzed

Organization of Methods

Write the Methods section in this order to ensure proper organization and make it easier for readers to understand how your study was carried out:

• Description of materials used, including site and sample
• Explanation of how materials were prepared
• Explanation of how measurements were made and calculations performed
• Explanation of statistical methods to analyze data

Tips for the Methods Section

• Organize description of preparations, measurements, and protocol chronologically
• List the Methods in the same order as they will appear in the Results section
• Material should be organized by topic from most to least important
• Headings can be used to separated different results; paragraphs are often used instead

Step 3: Write the Results

This section responds to the question  “What did you find?”  Only the direct results of  your  research should be presented here, not any results from other studies. This is essentially an analysis of the data explained in sentence form so that it is easier to read and put into context.

The Results section should include:

• Findings presented in the same order as in the Methods section
• Data presented in tables, charts, graphs, and other figures (placed among research text or on a separate page)
• Reports on data collection, recruitment, and/or participants
• Data that corresponds to the central research question(s)
• Secondary findings (secondary outcomes, subgroup analyses, etc.)

Organization of Results

Write the Results in the same order as you wrote your Methods. One trusted method of writing the results is addressing specific research questions presented in the figures. Within each research question, present the type of data that addresses that research question.

Sample research question asked in a survey:

“What do hospital patients over age 55 think about postoperative care?”

Present this answer as a statement based on the data:

“Hospital patients over the age of 55 were 30% more likely to report negative experiences after postoperative care (M=83; see Fig. 1).”

Elaborate on this finding with secondary information included in the same paragraph:

“The most common negative issues reported were inattention by nurses, lack of proper medicine and a prolonged waiting period for personal issues ((P>12), (W>13), and (D>10); see Fig. 3).”

Caption your figures with the same method, using the data and research question to create phrases that give context to the data:

“Figure 1: Attitudes towards postoperative care in patients over the age of 55.”

Grammar Guidelines for Results

• When referencing figures, use the present tense; when discussing events of the experiment/study, use past tense
• Passive or active voice are generally acceptable—but consistency is most important. (Read articles from target journal).
• Cite the figure or table every time you reference it, just as you would another text.

Dos and Don’ts for Results

• Limit your results to only those that address your research questions; return to the Results section later after you have completed the Introduction and remove less relevant information.
• Indicate the statistical tests used with all relevant parameters. E.g., mean and standard deviation (SD): 44% (±3); median and interpercentile range: 7 years (4.5 to 9.5 years).
• Use mean and standard deviation to report normally distributed data.
• Use median and interpercentile range to report skewed data.
• For numbers, use two significant digits unless more precision is necessary (2.08, not 2.07856444).
• Never use percentages for very small samples. E.g., “one out of two” should not be replaced by 50%.

Step 4: Write the Discussion/Conclusion

This section responds to the question  “What do the results mean?”  This section is easy to write, but difficult to write well. It requires more than a simply analysis—you have to interpret and “sell” your data to the journal and researchers, explaining just how important your findings are. In fact, many manuscripts are rejected because the Discussion section is weak.

The Discussion and Conclusion are often considered to be part of the same section, but the Conclusion is sometimes considered a separate section. At any rate, the Conclusion will be a very short and clear justification of your work or suggestion for future studies.

In the Discussion Section you should:

• Critique your study—be honest about the effectiveness of your design; suggest modifications and improvement.
• Answer this question: “Did your study contribute to knowledge in the field or not?”
• Discuss the impact of this research on related research within the domain

Pre-writing Questions to Answer for the Discussion:

• How do these results relate to the original question or objectives outlined in the Introduction section?
• Do the data support your hypothesis?
• Are your results consistent with what other investigators have reported?
• Discuss weaknesses and discrepancies. If your results were unexpected, try to explain why
• Is there another way to interpret your results?
• What further research would be necessary to answer the questions raised by your results?

Organization of the Discussion Section

The Discussion section is more open than the Results and Methods section, but you should always focus first on what is MOST important and then move to what is less important to your research problem. Divide the analysis of results by paragraph and do not combine unrelated datasets in one paragraph

• The first paragraph/part should summarize the process, the results, and the overall purpose of this study.
• The second paragraph/part should answer questions about the limitations and potential flaws or shortcomings of this study (e.g., the “failure to reveal clear relationships between samples or groups”). Assesses which of the results are most useful in answering the research question.
• The third paragraph should focus on the successes of the study and highlight which method or approach yielded the best results or those most closely hypothesized. You can also compare the results of different methods and assess which was more fruitful and why.
• In subsequent paragraphs, discuss the implications of this research and compare it to the results of other studies. This is the other section (in addition to the Introduction) where you can cite related studies to show how your study compares.

The Conclusion paragraph offers you a chance to briefly show how your work advances the field from the present state of knowledge. It adds a sort of exclamation point at the end of your paper and makes it more memorable as well.

Add a justification for your work here as well as indicate extensions and wider implications, as well as suggest future studies/experiments and point out any work that is currently ongoing. Do not simply repeat the Introduction or abstract here—extend the claims or questions raised in these sections.

Dos and Don’ts for Discussion/Conclusion

• Don’t be TOO broad about the impact of this research—set some limitations.
• Don’t include new terms or ideas in this section—they should be presented in the Introduction.
• Use specific expressions: instead of “higher temperature” write “41ºC”; instead of “at a lower rate” write “0.7% less”; instead of “highly significant” write “p<0.001.”

Step 5: Write the Introduction

The  Introduction section might be the most important section of the body of your paper—it comes first and introduces what you will be doing, telling readers why your work is important.

A good introduction should:

• Establish the context of the work
• State the purpose of the work in the form of a hypothesis, question, or problem investigated
• Give aims and rationale for your approach

Pre-writing questions to answer for the Introduction

• What is the problem to be solved? (background and problem)
• What do we know about this problem? (literature)
• Are there any existing solutions? (literature)
• What are the limitations or gaps in knowledge of existing solutions?
• What do you hope to achieve with this study? (hypothesis/statement of purpose)

Organization of the Introduction

• Background information
• Motivations
• Key primary literature
• Hypothesis/research problem investigated
• Approaches and rationale

Improving Quality of Writing

In order to write an effective research paper, authors need to know what areas of their writing to improve, and this includes avoiding grammar and style errors. Among the top writing errors we see at  Wordvice  are the following:

• Article and Determiner Misuses
• Nominalization and Wordiness
• Usage of Past and Present Tense

Receiving Language Editing Before Submission

After you are finished writing your Results section and have polished the rest of your research paper, be sure to submit your manuscript to an English proofreading service and paper editing service  before delivering it to journal editors for publication. And learn more about the  editing process  to determine which kind of revision your paper needs.

Wordvice Resources

• How to Write a Research Paper Introduction
• Writing the Results Section of a Research Paper
• Which Verb Tenses to Use in a Research Paper
• How to Write an Abstract for a Research Paper
• How to Write a Research Paper Title
• Useful Phrases for Academic Writing
• Common Transition Terms in Academic Papers
• Active and Passive Voice in Research Papers
• 100+ Verbs That Will Make Your Research Writing Amazing
• Tips for Paraphrasing in Research Papers

Related Resources

• Springer Online Research Resources  (Springer)
• ACSESS Digital L ibrary (ASA, CSSA, and SSSA publications for reference)  (ACSESS Digital Library)

Lecture Research Paper Reference

Yoon S-R, Kim SH, Lee H-W, Ha J-H (2017) A novel method to rapidly distinguish the geographical origin of traditional fermented-salted vegetables by mass fingerprinting. PLoS ONE 12(11): e0188217.

Have a language expert improve your writing

Run a free plagiarism check in 10 minutes, generate accurate citations for free.

• Knowledge Base
• Research paper

How to Create a Structured Research Paper Outline | Example

Published on August 7, 2022 by Courtney Gahan . Revised on August 15, 2023.

A research paper outline is a useful tool to aid in the writing process , providing a structure to follow with all information to be included in the paper clearly organized.

A quality outline can make writing your research paper more efficient by helping to:

• Organize your thoughts
• Understand the flow of information and how ideas are related
• Ensure nothing is forgotten

A research paper outline can also give your teacher an early idea of the final product.

Table of contents

Research paper outline example, how to write a research paper outline, formatting your research paper outline, language in research paper outlines.

• Definition of measles
• Rise in cases in recent years in places the disease was previously eliminated or had very low rates of infection
• Figures: Number of cases per year on average, number in recent years. Relate to immunization
• Symptoms and timeframes of disease
• Risk of fatality, including statistics
• How measles is spread
• Immunization procedures in different regions
• Different regions, focusing on the arguments from those against immunization
• Immunization figures in affected regions
• High number of cases in non-immunizing regions
• Illnesses that can result from measles virus
• Fatal cases of other illnesses after patient contracted measles
• Summary of arguments of different groups
• Summary of figures and relationship with recent immunization debate
• Which side of the argument appears to be correct?

Here's why students love Scribbr's proofreading services

Discover proofreading & editing

Follow these steps to start your research paper outline:

• Decide on the subject of the paper
• Write down all the ideas you want to include or discuss
• Organize related ideas into sub-groups
• Arrange your ideas into a hierarchy: What should the reader learn first? What is most important? Which idea will help end your paper most effectively?
• Create headings and subheadings that are effective
• Format the outline in either alphanumeric, full-sentence or decimal format

There are three different kinds of research paper outline: alphanumeric, full-sentence and decimal outlines. The differences relate to formatting and style of writing.

• Alphanumeric
• Full-sentence

An alphanumeric outline is most commonly used. It uses Roman numerals, capitalized letters, arabic numerals, lowercase letters to organize the flow of information. Text is written with short notes rather than full sentences.

• Sub-point of sub-point 1

Essentially the same as the alphanumeric outline, but with the text written in full sentences rather than short points.

• Additional sub-point to conclude discussion of point of evidence introduced in point A

A decimal outline is similar in format to the alphanumeric outline, but with a different numbering system: 1, 1.1, 1.2, etc. Text is written as short notes rather than full sentences.

• 1.1.1 Sub-point of first point
• 1.1.2 Sub-point of first point
• 1.2 Second point

To write an effective research paper outline, it is important to pay attention to language. This is especially important if it is one you will show to your teacher or be assessed on.

There are four main considerations: parallelism, coordination, subordination and division.

Parallelism: Be consistent with grammatical form

Parallel structure or parallelism is the repetition of a particular grammatical form within a sentence, or in this case, between points and sub-points. This simply means that if the first point is a verb , the sub-point should also be a verb.

Example of parallelism:

• Include different regions, focusing on the different arguments from those against immunization

Coordination: Be aware of each point’s weight

Your chosen subheadings should hold the same significance as each other, as should all first sub-points, secondary sub-points, and so on.

Example of coordination:

• Include immunization figures in affected regions
• Illnesses that can result from the measles virus

Subordination: Work from general to specific

Subordination refers to the separation of general points from specific. Your main headings should be quite general, and each level of sub-point should become more specific.

Example of subordination:

Division: break information into sub-points.

Your headings should be divided into two or more subsections. There is no limit to how many subsections you can include under each heading, but keep in mind that the information will be structured into a paragraph during the writing stage, so you should not go overboard with the number of sub-points.

Ready to start writing or looking for guidance on a different step in the process? Read our step-by-step guide on how to write a research paper .

Cite this Scribbr article

If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.

Gahan, C. (2023, August 15). How to Create a Structured Research Paper Outline | Example. Scribbr. Retrieved September 5, 2023, from https://www.scribbr.com/research-paper/outline/

Is this article helpful?

Courtney Gahan

Other students also liked, research paper format | apa, mla, & chicago templates, writing a research paper introduction | step-by-step guide, writing a research paper conclusion | step-by-step guide.

Courtney Gahan (Scribbr Team)

Thanks for reading! Hope you found this article helpful. If anything is still unclear, or if you didn’t find what you were looking for here, leave a comment and we’ll see if we can help.

Still have questions?

What is your plagiarism score.

• Utility Menu

26158766f7f76c0d163cbc4d15ae3f59

• Questions about Expos?
• Writing Support for Instructors

Your thesis is the central claim in your essay—your main insight or idea about your source or topic. Your thesis should appear early in an academic essay, followed by a logically constructed argument that supports this central claim. A strong thesis is arguable, which means a thoughtful reader could disagree with it and therefore needs your careful analysis of the evidence to understand how you arrived at this claim. You arrive at your thesis by examining and analyzing the evidence available to you, which might be text or other types of source material.

A thesis will generally respond to an analytical question or pose a solution to a problem that you have framed for your readers (and for yourself). When you frame that question or problem for your readers, you are telling them what is at stake in your argument—why your question matters and why they should care about the answer . If you can explain to your readers why a question or problem is worth addressing, then they will understand why it’s worth reading an essay that develops your thesis—and you will understand why it’s worth writing that essay.

A strong thesis will be arguable rather than descriptive , and it will be the right scope for the essay you are writing. If your thesis is descriptive, then you will not need to convince your readers of anything—you will be naming or summarizing something your readers can already see for themselves. If your thesis is too narrow, you won’t be able to explore your topic in enough depth to say something interesting about it. If your thesis is too broad, you may not be able to support it with evidence from the available sources.

When you are writing an essay for a course assignment, you should make sure you understand what type of claim you are being asked to make. Many of your assignments will be asking you to make analytical claims , which are based on interpretation of facts, data, or sources.

Some of your assignments may ask you to make normative claims. Normative claims are claims of value or evaluation rather than fact—claims about how things should be rather than how they are. A normative claim makes the case for the importance of something, the action that should be taken, or the way the world should be. When you are asked to write a policy memo, a proposal, or an essay based on your own opinion, you will be making normative claims.

Here are some examples of possible thesis statements for a student's analysis of the article “The Case Against Perfection” by Professor Michael Sandel.  

Descriptive thesis (not arguable)  

While Sandel argues that pursuing perfection through genetic engineering would decrease our sense of humility, he claims that the sense of solidarity we would lose is also important.

This thesis summarizes several points in Sandel’s argument, but it does not make a claim about how we should understand his argument. A reader who read Sandel’s argument would not also need to read an essay based on this descriptive thesis.  

Broad thesis (arguable, but difficult to support with evidence)  

Michael Sandel’s arguments about genetic engineering do not take into consideration all the relevant issues.

This is an arguable claim because it would be possible to argue against it by saying that Michael Sandel’s arguments do take all of the relevant issues into consideration. But the claim is too broad. Because the thesis does not specify which “issues” it is focused on—or why it matters if they are considered—readers won’t know what the rest of the essay will argue, and the writer won’t know what to focus on. If there is a particular issue that Sandel does not address, then a more specific version of the thesis would include that issue—hand an explanation of why it is important.  

Arguable thesis with analytical claim  

While Sandel argues persuasively that our instinct to “remake” (54) ourselves into something ever more perfect is a problem, his belief that we can always draw a line between what is medically necessary and what makes us simply “better than well” (51) is less convincing.

This is an arguable analytical claim. To argue for this claim, the essay writer will need to show how evidence from the article itself points to this interpretation. It’s also a reasonable scope for a thesis because it can be supported with evidence available in the text and is neither too broad nor too narrow.  

Arguable thesis with normative claim  

Given Sandel’s argument against genetic enhancement, we should not allow parents to decide on using Human Growth Hormone for their children.

This thesis tells us what we should do about a particular issue discussed in Sandel’s article, but it does not tell us how we should understand Sandel’s argument.  

Questions to ask about your thesis  

• Is the thesis truly arguable? Does it speak to a genuine dilemma in the source, or would most readers automatically agree with it?  
• Is the thesis too obvious? Again, would most or all readers agree with it without needing to see your argument?  
• Is the thesis complex enough to require a whole essay's worth of argument?  
• Is the thesis supportable with evidence from the text rather than with generalizations or outside research?  
• Would anyone want to read a paper in which this thesis was developed? That is, can you explain what this paper is adding to our understanding of a problem, question, or topic?
• Tips for Reading an Assignment Prompt
• Asking Analytical Questions
• Introductions
• What Do Introductions Across the Disciplines Have in Common?
• Anatomy of a Body Paragraph
• Transitions
• Tips for Organizing Your Essay
• Counterargument
• Conclusions
• Strategies for Essay Writing: Downloadable PDFs
• Brief Guides to Writing in the Disciplines

Quick Links

• Schedule an Appointment
• English Grammar and Language Tutor
• Harvard Guide to Using Sources
• Departmental Writing Fellows
• Writing Advice: The Harvard Writing Tutor Blog

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open Access
• Published: 30 August 2023

Champion-level drone racing using deep reinforcement learning

• Elia Kaufmann   ORCID: orcid.org/0000-0001-6094-5901 1 ,
• Leonard Bauersfeld 1 ,
• Antonio Loquercio   ORCID: orcid.org/0000-0002-8410-3933 1 ,
• Matthias Müller 2 ,
• Vladlen Koltun 3 &
• Davide Scaramuzza   ORCID: orcid.org/0000-0002-3831-6778 1  

Nature volume  620 ,  pages 982–987 ( 2023 ) Cite this article

49k Accesses

1098 Altmetric

Metrics details

• Aerospace engineering
• Computer science
• Electrical and electronic engineering
• Mechanical engineering

First-person view (FPV) drone racing is a televised sport in which professional competitors pilot high-speed aircraft through a 3D circuit. Each pilot sees the environment from the perspective of their drone by means of video streamed from an onboard camera. Reaching the level of professional pilots with an autonomous drone is challenging because the robot needs to fly at its physical limits while estimating its speed and location in the circuit exclusively from onboard sensors 1 . Here we introduce Swift, an autonomous system that can race physical vehicles at the level of the human world champions. The system combines deep reinforcement learning (RL) in simulation with data collected in the physical world. Swift competed against three human champions, including the world champions of two international leagues, in real-world head-to-head races. Swift won several races against each of the human champions and demonstrated the fastest recorded race time. This work represents a milestone for mobile robotics and machine intelligence 2 , which may inspire the deployment of hybrid learning-based solutions in other physical systems.

Deep RL 3 has enabled some recent advances in artificial intelligence. Policies trained with deep RL have outperformed humans in complex competitive games, including Atari 4 , 5 , 6 , Go 5 , 7 , 8 , 9 , chess 5 , 9 , StarCraft 10 , Dota 2 (ref.  11 ) and Gran Turismo 12 , 13 . These impressive demonstrations of the capabilities of machine intelligence have primarily been limited to simulation and board-game environments, which support policy search in an exact replica of the testing conditions. Overcoming this limitation and demonstrating champion-level performance in physical competitions is a long-standing problem in autonomous mobile robotics and artificial intelligence 14 , 15 , 16 .

FPV drone racing is a televised sport in which highly trained human pilots push aerial vehicles to their physical limits in high-speed agile manoeuvres (Fig. 1a ). The vehicles used in FPV racing are quadcopters, which are among the most agile machines ever built (Fig. 1b ). During a race, the vehicles exert forces that surpass their own weight by a factor of five or more, reaching speeds of more than 100 km h −1 and accelerations several times that of gravity, even in confined spaces. Each vehicle is remotely controlled by a human pilot who wears a headset showing a video stream from an onboard camera, creating an immersive ‘first-person-view’ experience (Fig. 1c ).

a , Swift (blue) races head-to-head against Alex Vanover, the 2019 Drone Racing League world champion (red). The track comprises seven square gates that must be passed in order in each lap. To win a race, a competitor has to complete three consecutive laps before its opponent. b , A close-up view of Swift, illuminated with blue LEDs, and a human-piloted drone, illuminated with red LEDs. The autonomous drones used in this work rely only on onboard sensory measurements, with no support from external infrastructure, such as motion-capture systems. c , From left to right: Thomas Bitmatta, Marvin Schaepper and Alex Vanover racing their drones through the track. Each pilot wears a headset that shows a video stream transmitted in real time from a camera aboard their aircraft. The headsets provide an immersive ‘first-person-view’ experience. c , Photo by Regina Sablotny.

Attempts to create autonomous systems that reach the performance of human pilots date back to the first autonomous drone racing competition in 2016 (ref.  17 ). A series of innovations followed, including the use of deep networks to identify the next gate location 18 , 19 , 20 , transfer of racing policies from simulation to reality 21 , 22 and accounting for uncertainty in perception 23 , 24 . The 2019 AlphaPilot autonomous drone racing competition showcased some of the best research in the field 25 . However, the first two teams still took almost twice as long as a professional human pilot to complete the track 26 , 27 . More recently, autonomous systems have begun to reach expert human performance 28 , 29 , 30 . However, these works rely on near-perfect state estimation provided by an external motion-capture system. This makes the comparison with human pilots unfair, as humans only have access to onboard observations from the drone.

In this article, we describe Swift, an autonomous system that can race a quadrotor at the level of human world champions using only onboard sensors and computation. Swift consists of two key modules: (1) a perception system that translates high-dimensional visual and inertial information into a low-dimensional representation and (2) a control policy that ingests the low-dimensional representation produced by the perception system and produces control commands.

The control policy is represented by a feedforward neural network and is trained in simulation using model-free on-policy deep RL 31 . To bridge discrepancies in sensing and dynamics between simulation and the physical world, we make use of non-parametric empirical noise models estimated from data collected on the physical system. These empirical noise models have proved to be instrumental for successful transfer of the control policy from simulation to reality.

We evaluate Swift on a physical track designed by a professional drone-racing pilot (Fig. 1a ). The track comprises seven square gates arranged in a volume of 30 × 30 × 8 m, forming a lap of 75 m in length. Swift raced this track against three human champions: Alex Vanover, the 2019 Drone Racing League world champion, Thomas Bitmatta, two-time MultiGP International Open World Cup champion, and Marvin Schaepper, three-time Swiss national champion. The quadrotors used by Swift and by the human pilots have the same weight, shape and propulsion. They are similar to drones used in international competitions.

The human pilots were given one week of practice on the race track. After this week of practice, each pilot competed against Swift in several head-to-head races (Fig. 1a,b ). In each head-to-head race, two drones (one controlled by a human pilot and one controlled by Swift) start from a podium. The race is set off by an acoustic signal. The first vehicle that completes three full laps through the track, passing all gates in the correct order in each lap, wins the race.

Swift won several races against each of the human pilots and achieved the fastest race time recorded during the events. Our work marks the first time, to our knowledge, that an autonomous mobile robot achieved world-champion-level performance in a real-world competitive sport.

The Swift system

Swift uses a combination of learning-based and traditional algorithms to map onboard sensory readings to control commands. This mapping comprises two parts: (1) an observation policy, which distils high-dimensional visual and inertial information into a task-specific low-dimensional encoding, and (2) a control policy that transforms the encoding into commands for the drone. A schematic overview of the system is shown in Fig. 2 .

Swift consists of two key modules: a perception system that translates visual and inertial information into a low-dimensional state observation and a control policy that maps this state observation to control commands. Control commands specify desired collective thrust and body rates, the same control modality that the human pilots use. a , The perception system consists of a VIO module that computes a metric estimate of the drone state from camera images and high-frequency measurements obtained by an inertial measurement unit (IMU). The VIO estimate is coupled with a neural network that detects the corners of racing gates in the image stream. The corner detections are mapped to a 3D pose and fused with the VIO estimate using a Kalman filter. b , We use model-free on-policy deep RL to train the control policy in simulation. During training, the policy maximizes a reward that combines progress towards the centre of the next racing gate with a perception objective to keep the next gate in the field of view of the camera. To transfer the racing policy from simulation to the physical world, we augment the simulation with data-driven residual models of the vehicle’s perception and dynamics. These residual models are identified from real-world experience collected on the race track. MLP, multilayer perceptron.

The observation policy consists of a visual–inertial estimator 32 , 33 that operates together with a gate detector 26 , which is a convolutional neural network that detects the racing gates in the onboard images. Detected gates are then used to estimate the global position and orientation of the drone along the race track. This is done by a camera-resectioning algorithm 34 in combination with a map of the track. The estimate of the global pose obtained from the gate detector is then combined with the estimate from the visual–inertial estimator by means of a Kalman filter, resulting in a more accurate representation of the robot’s state. The control policy, represented by a two-layer perceptron, maps the output of the Kalman filter to control commands for the aircraft. The policy is trained using on-policy model-free deep RL 31 in simulation. During training, the policy maximizes a reward that combines progress towards the next racing gate 35 with a perception objective that rewards keeping the next gate in the field of view of the camera. Seeing the next gate is rewarded because it increases the accuracy of pose estimation.

Optimizing a policy purely in simulation yields poor performance on physical hardware if the discrepancies between simulation and reality are not mitigated. The discrepancies are caused primarily by two factors: (1) the difference between simulated and real dynamics and (2) the noisy estimation of the robot’s state by the observation policy when provided with real sensory data. We mitigate these discrepancies by collecting a small amount of data in the real world and using this data to increase the realism of the simulator.

Specifically, we record onboard sensory observations from the robot together with highly accurate pose estimates from a motion-capture system while the drone is racing through the track. During this data-collection phase, the robot is controlled by a policy trained in simulation that operates on the pose estimates provided by the motion-capture system. The recorded data allow to identify the characteristic failure modes of perception and dynamics observed through the race track. These intricacies of failing perception and unmodelled dynamics are dependent on the environment, platform, track and sensors. The perception and dynamics residuals are modelled using Gaussian processes 36 and k -nearest-neighbour regression, respectively. The motivation behind this choice is that we empirically found perception residuals to be stochastic and dynamics residuals to be largely deterministic (Extended Data Fig. 1 ). These residual models are integrated into the simulation and the racing policy is fine-tuned in this augmented simulation. This approach is related to the empirical actuator models used for simulation-to-reality transfer in ref.  37 but further incorporates empirical modelling of the perception system and also accounts for the stochasticity in the estimate of the platform state.

We ablate each component of Swift in controlled experiments reported in the extended data. Also, we compare against recent work that tackles the task of autonomous drone racing with traditional methods, including trajectory planning and model predictive control (MPC). Although such approaches achieve comparable or even superior performance to our approach in idealized conditions, such as simplified dynamics and perfect knowledge of the robot’s state, their performance collapses when their assumptions are violated. We find that approaches that rely on precomputed paths 28 , 29 are particularly sensitive to noisy perception and dynamics. No traditional method has achieved competitive lap times compared with Swift or human world champions, even when provided with highly accurate state estimation from a motion-capture system. Detailed analysis is provided in the extended data.

The drone races take place on a track designed by an external world-class FPV pilot. The track features characteristic and challenging manoeuvres, such as a Split-S (Figs. 1a (top-right corner) and 4d ). Pilots are allowed to continue racing even after a crash, provided their vehicle is still able to fly. If both drones crash and cannot complete the track, the drone that proceeded farther along the track wins.

As shown in Fig. 3b , Swift wins 5 out of 9 races against A. Vanover, 4 out of 7 races against T. Bitmatta and 6 out of 9 races against M. Schaepper. Out of the 10 losses recorded for Swift, 40% were because of a collision with the opponent, 40% because of collision with a gate and 20% because of the drone being slower than the human pilot. Overall, Swift wins most races against each human pilot. Swift also achieves the fastest race time recorded, with a lead of half a second over the best time clocked by a human pilot (A. Vanover).

a , Lap-time results. We compare Swift against the human pilots in time-trial races. Lap times indicate best single lap times and best average times achieved in a heat of three consecutive laps. The reported statistics are computed over a dataset recorded during one week on the race track, which corresponds to 483 (115) data points for Swift, 331 (221) for A. Vanover, 469 (338) for T. Bitmatta and 345 (202) for M. Schaepper. The first number is the number of single laps and the second is the number of three consecutive laps. The dark points in each distribution correspond to laps flown in race conditions. b , Head-to-head results. We report the number of head-to-head races flown by each pilot, the number of wins and losses, as well as the win ratio.

Figure 4 and Extended Data Table 1d provide an analysis of the fastest lap flown by Swift and each human pilot. Although Swift is globally faster than all human pilots, it is not faster on all individual segments of the track (Extended Data Table 1 ). Swift is consistently faster at the start and in tight turns such as the split S. At the start, Swift has a lower reaction time, taking off from the podium, on average, 120 ms before human pilots. Also, it accelerates faster and reaches higher speeds going into the first gate (Extended Data Table 1d , segment 1). In sharp turns, as shown in Fig. 4c,d , Swift finds tighter manoeuvres. One hypothesis is that Swift optimizes trajectories on a longer timescale than human pilots. It is known that model-free RL can optimize long-term rewards through a value function 38 . Conversely, human pilots plan their motion on a shorter timescale, up to one gate into the future 39 . This is apparent, for example in the split S (Fig. 4b,d ), for which human pilots are faster in the beginning and at the end of the manoeuvre, but slower overall (Extended Data Table 1d , segment 3). Also, human pilots orient the aircraft to face the next gate earlier than Swift does (Fig. 4c,d ). We propose that human pilots are accustomed to keeping the upcoming gate in view, whereas Swift has learned to execute some manoeuvres while relying on other cues, such as inertial data and visual odometry against features in the surrounding environments. Overall, averaged over the entire track, the autonomous drone achieves the highest average speed, finds the shortest racing line and manages to maintain the aircraft closer to its actuation limits throughout the race, as indicated by the average thrust and power drawn (Extended Data Table 1d ).

a , Comparison of the fastest race of each pilot, illustrated by the time behind Swift. The time difference from the autonomous drone is computed as the time since it passed the same position on the track. Although Swift is globally faster than all human pilots, it is not necessarily faster on all individual segments of the track. b , Visualization of where the human pilots are faster (red) and slower (blue) compared with the autonomous drone. Swift is consistently faster at the start and in tight turns, such as the split S. c , Analysis of the manoeuvre after gate 2. Swift in blue, Vanover in red. Swift gains time against human pilots in this segment as it executes a tighter turn while maintaining comparable speed. d , Analysis of the split S manoeuvre. Swift in blue, Vanover in red. The split S is the most challenging segment in the race track, requiring a carefully coordinated roll and pitch motion that yields a descending half-loop through the two gates. Swift gains time against human pilots on this segment as it executes a tighter turn with less overshoot. e , Illustration of track segments used for analysis. Segment 1 is traversed once at the start, whereas segments 2–4 are traversed in each lap (three times over the course of a race).

We also compare the performance of Swift and the human champions in time trials (Fig. 3a ). In a time trial, a single pilot races the track, with the number of laps left to the discretion of the pilot. We accumulate time-trial data from the practice week and the races, including training runs (Fig. 3a , coloured) and laps flown in race conditions (Fig. 3a , black). For each contestant, we use more than 300 laps for computing statistics. The autonomous drone more consistently pushes for fast lap times, exhibiting lower mean and variance. Conversely, human pilots decide whether to push for speed on a lap-by-lap basis, yielding higher mean and variance in lap times, both during training and in the races. The ability to adapt the flight strategy allows human pilots to maintain a slower pace if they identify that they have a clear lead, so as to reduce the risk of a crash. The autonomous drone is unaware of its opponent and pushes for fastest expected completion time no matter what, potentially risking too much when in the lead and too little when trailing behind 40 .

FPV drone racing requires real-time decision-making based on noisy and incomplete sensory input from the physical environment. We have presented an autonomous physical system that achieves champion-level performance in this sport, reaching—and at times exceeding—the performance of human world champions. Our system has certain structural advantages over the human pilots. First, it makes use of inertial data from an onboard inertial measurement unit 32 . This is akin to the human vestibular system 41 , which is not used by the human pilots because they are not physically in the aircraft and do not feel the accelerations acting on it. Second, our system benefits from lower sensorimotor latency (40 ms for Swift versus an average of 220 ms for expert human pilots 39 ). On the other hand, the limited refresh rate of the camera used by Swift (30 Hz) can be considered a structural advantage for human pilots, whose cameras’ refresh rate is four times as fast (120 Hz), improving their reaction time 42 .

Human pilots are impressively robust: they can crash at full speed, and—if the hardware still functions—carry on flying and complete the track. Swift was not trained to recover after a crash. Human pilots are also robust to changes in environmental conditions, such as illumination, which can markedly alter the appearance of the track. By contrast, Swift’s perception system assumes that the appearance of the environment is consistent with what was observed during training. If this assumption fails, the system can fail. Robustness to appearance changes can be provided by training the gate detector and the residual observation model in a diverse set of conditions. Addressing these limitations could enable applying the presented approach in autonomous drone racing competitions in which access to the environment and the drone is limited 25 .

Notwithstanding the remaining limitations and the work ahead, the attainment by an autonomous mobile robot of world-champion-level performance in a popular physical sport is a milestone for robotics and machine intelligence. This work may inspire the deployment of hybrid learning-based solutions in other physical systems, such as autonomous ground vehicles, aircraft and personal robots, across a broad range of applications.

Quadrotor simulation

Quadrotor dynamics.

To enable large-scale training, we use a high-fidelity simulation of the quadrotor dynamics. This section briefly explains the simulation. The dynamics of the vehicle can be written as

in which ⊙ represents quaternion rotation, \({{\bf{p}}}_{{\mathcal{W}}{\mathcal{B}}},{{\bf{q}}}_{{\mathcal{W}}{\mathcal{B}}},{{\bf{v}}}_{{\mathcal{W}}}\) and \({{\boldsymbol{\omega }}}_{{\mathcal{B}}}\) denote the position, attitude quaternion, inertial velocity and body rates of the quadcopter, respectively. The motor time constant is k mot and the motor speeds Ω and Ω ss are the actual and steady-state motor speeds, respectively. The matrix J  is the inertia of the quadcopter and \({{\bf{g}}}_{{\mathcal{W}}}\) denotes the gravity vector. Two forces act on the quadrotor: the lift force f prop generated by the propellers and an aerodynamic force f aero that aggregates all other forces, such as aerodynamic drag, dynamic lift and induced drag. The torque is modelled as a sum of four components: the torque generated by the individual propeller thrusts τ prop , the yaw torque τ mot generated by a change in motor speed, an aerodynamic torque τ aero that accounts for various aerodynamic effects such as blade flapping and an inertial term τ iner . The individual components are given as

in which r P, i is the location of propeller i , expressed in the body frame, and f i and τ i are the forces and torques, respectively, generated by the i th propeller. The axis of rotation of the i th motor is denoted by ζ i , the combined inertia of the motor and propeller is J m+p and the derivative of the i th motor speed is \({\dot{\Omega }}_{i}\) . The individual propellers are modelled using a commonly used quadratic model, which assumes that the lift force and drag torque are proportional to the square of the propeller speed Ω i :

in which c l and c d denote the propeller lift and drag coefficients, respectively.

Aerodynamic forces and torques

The aerodynamic forces and torques are difficult to model with a first-principles approach. We thus use a data-driven model 43 . To maintain the low computational complexity required for large-scale RL training, a grey-box polynomial model is used rather than a neural network. The aerodynamic effects are assumed to primarily depend on the velocity \({{\bf{v}}}_{{\mathcal{B}}}\) (in the body frame) and the average squared motor speed \(\overline{{\Omega }^{2}}\) . The aerodynamic forces f x , f y and f z and torques τ x , τ y and τ z are estimated in the body frame. The quantities v x , v y and v z denote the three axial velocity components (in the body frame) and v x y denotes the speed in the ( x ,  y ) plane of the quadrotor. On the basis of insights from the underlying physical processes, linear and quadratic combinations of the individual terms are selected. For readability, the coefficients multiplying each summand have been omitted:

The respective coefficients are then identified from real-world flight data, in which motion capture is used to provide ground-truth forces and torque measurements. We use data from the race track, allowing the dynamics model to fit the track. This is akin to the human pilots’ training for days or weeks before the race on the specific track that they will be racing. In our case, the human pilots are given a week of practice on the same track ahead of the competition.

Betaflight low-level controller

To control the quadrotor, the neural network outputs collective thrust and body rates. This control signal is known to combine high agility with good robustness to simulation-to-reality transfer 44 . The predicted collective thrust and body rates are then processed by an onboard low-level controller that computes individual motor commands, which are subsequently translated into analogue voltage signals through an electronic speed controller (ESC) that controls the motors. On the physical vehicle, this low-level proportional–integral–derivative (PID) controller and ESC are implemented using the open-source Betaflight and BLHeli32 firmware 45 . In simulation, we use an accurate model of both the low-level controller and the motor speed controller.

Because the Betaflight PID controller has been optimized for human-piloted flight, it exhibits some peculiarities, which the simulation correctly captures: the reference for the D-term is constantly zero (pure damping), the I-term gets reset when the throttle is cut and, under motor thrust saturation, the body rate control is assigned priority (proportional downscaling of all motor signals to avoid saturation). The gains of the controller used for simulation have been identified from the detailed logs of the Betaflight controller’s internal states. The simulation can predict the individual motor commands with less than 1% error.

Battery model and ESC

The low-level controller converts the individual motor commands into a pulse-width modulation (PWM) signal and sends it to the ESC, which controls the motors. Because the ESC does not perform closed-loop control of the motor speeds, the steady-state motor speed Ω i ,ss for a given PWM motor command cmd i is a function of the battery voltage. Our simulation thus models the battery voltage using a grey-box battery model 46 that simulates the voltage based on instantaneous power consumption P mot :

The battery model 46 then simulates the battery voltage based on this power demand. Given the battery voltage U bat and the individual motor command u cmd, i , we use the mapping (again omitting the coefficients multiplying each summand)

to calculate the corresponding steady-state motor speed Ω i ,ss required for the dynamics simulation in equation ( 1 ). The coefficients have been identified from Betaflight logs containing measurements of all involved quantities. Together with the model of the low-level controller, this enables the simulator to correctly translate an action in the form of collective thrust and body rates to desired motor speeds Ω ss in equation ( 1 ).

Policy training

We train deep neural control policies that directly map observations o t in the form of platform state and next gate observation to control actions u t in the form of mass-normalized collective thrust and body rates 44 . The control policies are trained using model-free RL in simulation.

Training algorithm

Training is performed using proximal policy optimization 31 . This actor-critic approach requires jointly optimizing two neural networks during training: the policy network, which maps observations to actions, and the value network, which serves as the ‘critic’ and evaluates actions taken by the policy. After training, only the policy network is deployed on the robot.

Observations, actions and rewards

An observation \({{\bf{o}}}_{t}\in {{\mathbb{R}}}^{31}\) obtained from the environment at time t consists of: (1) an estimate of the current robot state; (2) the relative pose of the next gate to be passed on the track layout; and (3) the action applied in the previous step. Specifically, the estimate of the robot state contains the position of the platform, its velocity and attitude represented by a rotation matrix, resulting in a vector in \({{\mathbb{R}}}^{15}\) . Although the simulation uses quaternions internally, we use a rotation matrix to represent attitude to avoid ambiguities 47 . The relative pose of the next gate is encoded by providing the relative position of the four gate corners with respect to the vehicle, resulting in a vector in \({{\mathbb{R}}}^{12}\) . All observations are normalized before being passed to the network. Because the value network is only used during training time, it can access privileged information about the environment that is not accessible to the policy 48 . This privileged information is concatenated with other inputs to the policy network and contains the exact position, orientation and velocity of the robot.

For each observation o t , the policy network produces an action \({{\bf{a}}}_{t}\in {{\mathbb{R}}}^{4}\) in the form of desired mass-normalized collective thrust and body rates.

We use a dense shaped reward formulation to learn the task of perception-aware autonomous drone racing. The reward r t at time step t is given by

in which r prog rewards progress towards the next gate 35 , r perc encodes perception awareness by adjusting the attitude of the vehicle such that the optical axis of the camera points towards the centre of the next gate, r cmd rewards smooth actions and r crash is a binary penalty that is only active when colliding with a gate or when the platform leaves a predefined bounding box. If r crash is triggered, the training episode ends.

Specifically, the reward terms are

in which \({d}_{t}^{{\rm{Gate}}}\) denotes the distance from the centre of mass of the vehicle to the centre of the next gate at time step t , δ cam represents the angle between the optical axis of the camera and the centre of the next gate and \({{\bf{a}}}_{t}^{\omega }\) are the commanded body rates. The hyperparameters λ 1 ,…,  λ 5 balance different terms (Extended Data Table 1a ).

Training details

Data collection is performed by simulating 100 agents in parallel that interact with the environment in episodes of 1,500 steps. At each environment reset, every agent is initialized at a random gate on the track, with bounded perturbation around a state previously observed when passing this gate. In contrast to previous work 44 , 49 , 50 , we do not perform randomization of the platform dynamics at training time. Instead, we perform fine-tuning based on real-world data. The training environment is implemented using TensorFlow Agents 51 . The policy network and the value network are both represented by two-layer perceptrons with 128 nodes in each layer and LeakyReLU activations with a negative slope of 0.2. Network parameters are optimized using the Adam optimizer with learning rate 3 × 10 −4 for both the policy network and the value network.

Policies are trained for a total of 1 × 10 8 environment interactions, which takes 50 min on a workstation (i9 12900K, RTX 3090, 32 GB RAM DDR5). Fine-tuning is performed for 2 × 10 7 environment interactions.

Residual model identification

We perform fine-tuning of the original policy based on a small amount of data collected in the real world. Specifically, we collect three full rollouts in the real world, corresponding to approximately 50 s of flight time. We fine-tune the policy by identifying residual observations and residual dynamics, which are then used for training in simulation. During this fine-tuning phase, only the weights of the control policy are updated, whereas the weights of the gate-detection network are kept constant.

Residual observation model

Navigating at high speeds results in substantial motion blur, which can lead to a loss of tracked visual features and severe drift in linear odometry estimates. We fine-tune policies with an odometry model that is identified from only a handful of trials recorded in the real world. To model the drift in odometry, we use Gaussian processes 36 , as they allow fitting a posterior distribution of odometry perturbations, from which we can sample temporally consistent realizations.

Specifically, the Gaussian process model fits residual position, velocity and attitude as a function of the ground-truth robot state. The observation residuals are identified by comparing the observed visual–inertial odometry (VIO) estimates during a real-world rollout with the ground-truth platform states, which are obtained from an external motion-tracking system.

We treat each dimension of the observation separately, effectively fitting a set of nine 1D Gaussian processes to the observation residuals. We use a mixture of radial basis function kernels

in which L  is the diagonal length scale matrix and σ f and σ n represent the data and prior noise variance, respectively, and z i and z j represent data features. The kernel hyperparameters are optimized by maximizing the log marginal likelihood. After kernel hyperparameter optimization, we sample new realizations from the posterior distribution that are then used during fine-tuning of the policy. Extended Data Fig. 1 illustrates the residual observations in position, velocity and attitude in real-world rollouts, as well as 100 sampled realizations from the Gaussian process model.

Residual dynamics model

We use a residual model to complement the simulated robot dynamics 52 . Specifically, we identify residual accelerations as a function of the platform state s and the commanded mass-normalized collective thrust c :

We use k -nearest neighbour regression with k  = 5. The size of the dataset used for residual dynamics model identification depends on the track layout and ranges between 800 and 1,000 samples for the track layout used in this work.

Gate detection

To correct for drift accumulated by the VIO pipeline, the gates are used as distinct landmarks for relative localization. Specifically, gates are detected in the onboard camera view by segmenting gate corners 26 . The greyscale images provided by the Intel RealSense Tracking Camera T265 are used as input images for the gate detector. The architecture of the segmentation network is a six-level U-Net 53 with (8, 16, 16, 16, 16, 16) convolutional filters of size (3, 3, 3, 5, 7, 7) per level and a final extra layer operating on the output of the U-Net containing 12 filters. As the activation function, LeakyReLU with α  = 0.01 is used. For deployment on the NVIDIA Jetson TX2, the network is ported to TensorRT. To optimize memory footprint and computation time, inference is performed in half-precision mode (FP16) and images are downsampled to size 384 × 384 before being fed to the network. One forward pass through the network takes 40 ms on the NVIDIA Jetson TX2.

VIO drift estimation

The odometry estimates from the VIO pipeline 54 exhibit substantial drift during high-speed flight. We use gate detection to stabilize the pose estimates produced by VIO. The gate detector outputs the coordinates of the corners of all visible gates. A relative pose is first estimated for all predicted gates using infinitesimal plane-based pose estimation (IPPE) 34 . Given this relative pose estimate, each gate observation is assigned to the closest gate in the known track layout, thus yielding a pose estimate for the drone.

Owing to the low frequency of the gate detections and the high quality of the VIO orientation estimate, we only refine the translational components of the VIO measurements. We estimate and correct for the drift of the VIO pipeline using a Kalman filter that estimates the translational drift p d (position offset) and its derivative, the drift velocity v d . The drift correction is performed by subtracting the estimated drift states p d and v d from the corresponding VIO estimates. The Kalman filter state x is given by \({\bf{x}}={[{{\bf{p}}}_{{\rm{d}}}^{\top },{{\bf{v}}}_{{\rm{d}}}^{\top }]}^{\top }\in {{\mathbb{R}}}^{6}\) .

The state x and covariance P  updates are given by:

On the basis of measurements, the process noise is set to σ pos  = 0.05 and σ vel  = 0.1. The filter state and covariance are initialized to zero. For each measurement z k (pose estimate from a gate detection), the predicted VIO drift \({{\bf{x}}}_{k}^{-}\) is corrected to the estimate \({{\bf{x}}}_{k}^{+}\) according to the Kalman filter equations:

in which K k  is the Kalman gain, R  is the measurement covariance and H k is the measurement matrix. If several gates have been detected in a single camera frame, all relative pose estimates are stacked and processed in the same Kalman filter update step. The main source of measurement error is the uncertainty in the gate-corner detection of the network. This error in the image plane results in a pose error when IPPE is applied. We opted for a sampling-based approach to estimate the pose error from the known average gate-corner-detection uncertainty. For each gate, the IPPE algorithm is applied to the nominal gate observation as well as to 20 perturbed gate-corner estimates. The resulting distribution of pose estimates is then used to approximate the measurement covariance R  of the gate observation.

Simulation results

Reaching champion-level performance in autonomous drone racing requires overcoming two challenges: imperfect perception and incomplete models of the system’s dynamics. In controlled experiments in simulation, we assess the robustness of our approach to both of these challenges. To this end, we evaluate performance in a racing task when deployed in four different settings. In setting (1), we simulate a simplistic quadrotor model with access to ground-truth state observations. In setting (2), we replace the ground-truth state observations with noisy observations identified from real-world flights. These noisy observations are generated by sampling one realization from the residual observation model and are independent of the perception awareness of the deployed controller. Settings (3) and (4) share the observation models with the previous two settings, respectively, but replace the simplistic dynamics model with more accurate aerodynamical simulation 43 . These four settings allow controlled assessment of the sensitivity of the approach to changes in the dynamics and the observation fidelity.

In all four settings, we benchmark our approach against the following baselines: zero-shot, domain randomization and time-optimal. The zero-shot baseline represents a learning-based racing policy 35 trained using model-free RL that is deployed zero-shot from the training domain to the test domain. The training domain of the policy is equal to experimental setting (1), that is, idealized dynamics and ground-truth observations. Domain randomization extends the learning strategy from the zero-shot baseline by randomizing observations and dynamics properties to increase robustness. The time-optimal baseline uses a precomputed time-optimal trajectory 28 that is tracked using an MPC controller. This approach has shown the best performance in comparison with other model-based methods for time-optimal flight 55 , 56 . The dynamics model used by the trajectory generation and the MPC controller matches the simulated dynamics of experimental setting (1).

Performance is assessed by evaluating the fastest lap time, the average and minimum observed gate margin of successfully passed gates and the percentage of track successfully completed. The gate margin metric measures the distance between the drone and the closest point on the gate when crossing the gate plane. A high gate margin indicates that the quadrotor passed close to the centre of the gate. Leaving a smaller gate margin can increase speed but can also increase the risk of collision or missing the gate. Any lap that results in a crash is not considered valid.

The results are summarized in Extended Data Table 1c . All approaches manage to successfully complete the task when deployed in idealized dynamics and ground-truth observations, with the time-optimal baseline yielding the lowest lap time. When deployed in settings that feature domain shift, either in the dynamics or the observations, the performance of all baselines collapses and none of the three baselines are able to complete even a single lap. This performance drop is exhibited by both learning-based and traditional approaches. By contrast, our approach, which features empirical models of dynamics and observation noise, succeeds in all deployment settings, with small increases in lap time.

The key feature that enables our approach to succeed across deployment regimes is the use of an empirical model of dynamics and observation noise, estimated from real-world data. A comparison between an approach that has access to such data and approaches that do not is not entirely fair. For that reason, we also benchmark the performance of all baseline approaches when having access to the same real-world data used by our approach. Specifically, we compare the performance in experimental setting (2), which features the idealized dynamics model but noisy perception. All baseline approaches are provided with the predictions of the same Gaussian process model that we use to characterize observation noise. The results are summarized in Extended Data Table 1b . All baselines benefit from the more realistic observations, yielding higher completion rates. Nevertheless, our approach is the only one that reliably completes the entire track. As well as the predictions of the observation noise model, our approach also takes into account the uncertainty of the model. For an in-depth comparison of the performance of RL versus optimal control in controlled experiments, we refer the reader to ref.  57 .

Fine-tuning for several iterations

We investigate the extent of variations in behaviour across iterations. The findings of our analysis reveal that subsequent fine-tuning operations result in negligible enhancements in performance and alterations in behaviour (Extended Data Fig. 2 ).

In the following, we provide more details on this investigation. We start by enumerating the fine-tuning steps to provide the necessary notation:

Train policy-0 in simulation.

Deploy policy-0 in the real world. The policy operates on ground-truth data from a motion-capture system.

Identify residuals observed by policy-0 in the real world.

Train policy-1 by fine-tuning policy-0 on the identified residuals.

Deploy policy-1 in the real world. The policy operates only on onboard sensory measurements.

Identify residuals observed by policy-1 in the real world.

Train policy-2 by fine-tuning policy-1 on the identified residuals.

We compare the performance of policy-1 and policy-2 in simulation after fine-tuning on their respective residuals. The results are illustrated in Extended Data Fig. 2 . We observe that the difference in distance from gate centres, which is a metric for the safety of the policy, is 0.09 ± 0.08 m. Furthermore, the difference in the time taken to complete a single lap is 0.02 ± 0.02 s. Note that this lap-time difference is substantially smaller than the difference between the single-lap completion times of Swift and the human pilots (0.16 s).

Drone hardware configuration

The quadrotors used by the human pilots and Swift have the same weight, shape and propulsion. The platform design is based on the Agilicious framework 58 . Each vehicle has a weight of 870 g and can produce a maximum static thrust of approximately 35 N, which results in a static thrust-to-weight ratio of 4.1. The base of each platform consists of an Armattan Chameleon 6″ main frame that is equipped with T-Motor Velox 2306 motors and 5″, three-bladed propellers. An NVIDIA Jetson TX2 accompanied by a Connect Tech Quasar carrier board provides the main compute resource for the autonomous drones, featuring a six-core CPU running at 2 GHz and a dedicated GPU with 256 CUDA cores running at 1.3 GHz. Although forward passes of the gate-detection network are performed on the GPU, the racing policy is evaluated on the CPU, with one inference pass taking 8 ms. The autonomous drones carry an Intel RealSense Tracking Camera T265 that provides VIO estimates 59 at 100 Hz that are fed by USB to the NVIDIA Jetson TX2. The human-piloted drones carry neither a Jetson computer nor a RealSense camera and are instead equipped with a corresponding ballast weight. Control commands in the form of collective thrust and body rates produced by the human pilots or Swift are sent to a commercial flight controller, which runs on an STM32 processor operating at 216 MHz. The flight controller is running Betaflight, an open-source flight-control software 45 .

Human pilot impressions

The following quotes convey the impressions of the three human champions who raced against Swift.

Alex Vanover :

These races will be decided at the split S, it is the most challenging part of the track.

This was the best race! I was so close to the autonomous drone, I could really feel the turbulence when trying to keep up with it.

Thomas Bitmatta :

The possibilities are endless, this is the start of something that could change the whole world. On the flip side, I’m a racer, I don’t want anything to be faster than me.

As you fly faster, you trade off precision for speed.

It’s inspiring to see the potential of what drones are actually capable of. Soon, the AI drone could even be used as a training tool to understand what would be possible.

Marvin Schaepper :

It feels different racing against a machine, because you know that the machine doesn’t get tired.

Research ethics

The study has been conducted in accordance with the Declaration of Helsinki. The study protocol is exempt from review by an ethics committee according to the rules and regulations of the University of Zurich, because no health-related data has been collected. The participants gave their written informed consent before participating in the study.

Data availability

All (other) data needed to evaluate the conclusions in the paper are present in the paper or the extended data. Motion-capture recordings of the race events with accompanying analysis code can be found in the file ‘racing_data.zip’ on Zenodo at https://doi.org/10.5281/zenodo.7955278 .

Code availability

Pseudocode for Swift detailing the training process and algorithms can be found in the file ‘pseudocode.zip’ on Zenodo at https://doi.org/10.5281/zenodo.7955278 . To safeguard against potential misuse, the full source code associated with this research will not be made publicly available.

De Wagter, C., Paredes-Vallés, F., Sheth, N. & de Croon, G. Learning fast in autonomous drone racing. Nat. Mach. Intell. 3 , 923 (2021).

Article   Google Scholar  

Hanover, D. et al. Autonomous drone racing: a survey. Preprint at https://arxiv.org/abs/2301.01755 (2023).

Sutton, R. S. & Barto, A. G. Reinforcement Learning: An Introduction (MIT Press, 2018).

Mnih, V. et al. Human-level control through deep reinforcement learning. Nature 518 , 529–533 (2015).

Article   ADS   CAS   PubMed   Google Scholar  

Schrittwieser, J. et al. Mastering Atari, Go, chess and shogi by planning with a learned model. Nature 588 , 604–609 (2020).

Ecoffet, A., Huizinga, J., Lehman, J., Stanley, K. O. & Clune, J. First return, then explore. Nature 590 , 580–586 (2021).

Silver, D. et al. Mastering the game of Go with deep neural networks and tree search. Nature 529 , 484–489 (2016).

Silver, D. et al. Mastering the game of Go without human knowledge. Nature 550 , 354–359 (2017).

Silver, D. et al. A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play. Science 362 , 1140–1144 (2018).

Article   ADS   MathSciNet   CAS   PubMed   MATH   Google Scholar  

Vinyals, O. et al. Grandmaster level in StarCraft II using multi-agent reinforcement learning. Nature 575 , 350–354 (2019).

Berner, C. et al. Dota 2 with large scale deep reinforcement learning. Preprint at https://arxiv.org/abs/1912.06680 (2019).

Fuchs, F., Song, Y., Kaufmann, E., Scaramuzza, D. & Dürr, P. Super-human performance in Gran Turismo Sport using deep reinforcement learning. IEEE Robot. Autom. Lett. 6 , 4257–4264 (2021).

Wurman, P. R. et al. Outracing champion Gran Turismo drivers with deep reinforcement learning. Nature 602 , 223–228 (2022).

Funke, J. et al. in Proc. 2012 IEEE Intelligent Vehicles Symposium 541–547 (IEEE, 2012).

Spielberg, N. A., Brown, M., Kapania, N. R., Kegelman, J. C. & Gerdes, J. C. Neural network vehicle models for high-performance automated driving. Sci. Robot. 4 , eaaw1975 (2019).

Article   PubMed   Google Scholar  

Won, D.-O., Müller, K.-R. & Lee, S.-W. An adaptive deep reinforcement learning framework enables curling robots with human-like performance in real-world conditions. Sci. Robot. 5 , eabb9764 (2020).

Moon, H., Sun, Y., Baltes, J. & Kim, S. J. The IROS 2016 competitions. IEEE Robot. Autom. Mag. 24 , 20–29 (2017).

Jung, S., Hwang, S., Shin, H. & Shim, D. H. Perception, guidance, and navigation for indoor autonomous drone racing using deep learning. IEEE Robot. Autom. Lett. 3 , 2539–2544 (2018).

Kaufmann, E. et al. in Proc. 2nd Conference on Robot Learning (CoRL) 133–145 (PMLR, 2018).

Zhang, D. & Doyle, D. D. in Proc. 2020 IEEE Aerospace Conference , 1–11 (IEEE, 2020).

Loquercio, A. et al. Deep drone racing: from simulation to reality with domain randomization. IEEE Trans. Robot. 36 , 1–14 (2019).

Loquercio, A. et al. Learning high-speed flight in the wild. Sci. Robot. 6 , eabg5810 (2021).

Kaufmann, E. et al. in Proc. 2019 International Conference on Robotics and Automation (ICRA) 690–696 (IEEE, 2019).

Li, S., van der Horst, E., Duernay, P., De Wagter, C. & de Croon, G. C. Visual model-predictive localization for computationally efficient autonomous racing of a 72-g drone. J. Field Robot. 37 , 667–692 (2020).

A.I. is flying drones (very, very slowly). https://www.nytimes.com/2019/03/26/technology/alphapilot-ai-drone-racing.html (2019).

Foehn, P. et al. AlphaPilot: autonomous drone racing. Auton. Robots 46 , 307–320 (2021).

Article   PubMed   PubMed Central   Google Scholar  

Wagter, C. D., Paredes-Vallé, F., Sheth, N. & de Croon, G. The sensing, state-estimation, and control behind the winning entry to the 2019 Artificial Intelligence Robotic Racing Competition. Field Robot. 2 , 1263–1290 (2022).

Foehn, P., Romero, A. & Scaramuzza, D. Time-optimal planning for quadrotor waypoint flight. Sci. Robot. 6 , eabh1221 (2021).

Romero, A., Sun, S., Foehn, P. & Scaramuzza, D. Model predictive contouring control for time-optimal quadrotor flight. IEEE Trans. Robot. 38 , 3340–3356 (2022).

Sun, S., Romero, A., Foehn, P., Kaufmann, E. & Scaramuzza, D. A comparative study of nonlinear MPC and differential-flatness-based control for quadrotor agile flight. IEEE Trans. Robot. 38 , 3357–3373 (2021).

Schulman, J., Wolski, F., Dhariwal, P., Radford, A. & Klimov, O. Proximal policy optimization algorithms. Preprint at https://arxiv.org/abs/1707.06347 (2017).

Scaramuzza, D. & Zhang, Z. Encyclopedia of Robotics (eds Ang, M., Khatib, O. & Siciliano, B.) 1–9 (Springer, 2019).

Huang, G. in Proc. 2019 International Conference on Robotics and Automation (ICRA) 9572–9582 (IEEE, 2019).

Collins, T. & Bartoli, A. Infinitesimal plane-based pose estimation. Int. J. Comput. Vis. 109 , 252–286 (2014).

Article   MathSciNet   MATH   Google Scholar  

Song, Y., Steinweg, M., Kaufmann, E. & Scaramuzza, D. in Proc. 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 1205–1212 (IEEE, 2021).

Williams, C. K. & Rasmussen, C. E. Gaussian Processes for Machine Learning (MIT Press, 2006).

Hwangbo, J. et al. Learning agile and dynamic motor skills for legged robots. Sci. Robot. 4 , eaau5872 (2019).

Hung, C.-C. et al. Optimizing agent behavior over long time scales by transporting value. Nat. Commun. 10 , 5223 (2019).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Pfeiffer, C. & Scaramuzza, D. Human-piloted drone racing: visual processing and control. IEEE Robot. Autom. Lett. 6 , 3467–3474 (2021).

Spica, R., Cristofalo, E., Wang, Z., Montijano, E. & Schwager, M. A real-time game theoretic planner for autonomous two-player drone racing. IEEE Trans. Robot. 36 , 1389–1403 (2020).

Day, B. L. & Fitzpatrick, R. C. The vestibular system. Curr. Biol. 15 , R583–R586 (2005).

Article   CAS   PubMed   Google Scholar  

Kim, J. et al. Esports arms race: latency and refresh rate for competitive gaming tasks. J. Vis. 19 , 218c (2019).

Bauersfeld, L., Kaufmann, E., Foehn, P., Sun, S. & Scaramuzza, D. in Proc. Robotics: Science and Systems XVII 42 (Robotics: Science and Systems Foundation, 2021).

Kaufmann, E., Bauersfeld, L. & Scaramuzza, D. in Proc. 2022 International Conference on Robotics and Automation (ICRA) 10504–10510 (IEEE, 2022).

The Betaflight Open Source Flight Controller Firmware Project. Betaflight. https://github.com/betaflight/betaflight (2022).

Bauersfeld, L. & Scaramuzza, D. Range, endurance, and optimal speed estimates for multicopters. IEEE Robot. Autom. Lett. 7 , 2953–2960 (2022).

Zhou, Y., Barnes, C., Lu, J., Yang, J. & Li, H. in Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 5745–5753 (IEEE, 2019).

Pinto, L., Andrychowicz, M., Welinder, P., Zaremba, W. & Abbeel, P. in Proc. Robotics: Science and Systems XIV (MIT Press Journals, 2018).

Molchanov, A. et al. in Proc. 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 59–66 (IEEE, 2019).

Andrychowicz, O. M. et al. Learning dexterous in-hand manipulation. Int. J. Robot. Res. 39 , 3–20 (2020).

Guadarrama, S. et al. TF-Agents: a library for reinforcement learning in TensorFlow. https://github.com/tensorflow/agents (2018).

Torrente, G., Kaufmann, E., Foehn, P. & Scaramuzza, D. Data-driven MPC for quadrotors. IEEE Robot. Autom. Lett. 6 , 3769–3776 (2021).

Ronneberger, O., Fischer, P. & Brox, T. in Proc. International Conference on Medical Image Computing and Computer-assisted Intervention 234–241 (Springer, 2015).

Intel RealSense T265 series product family. https://www.intelrealsense.com/wp-content/uploads/2019/09/Intel_RealSense_Tracking_Camera_Datasheet_Rev004_release.pdf (2019).

Ryou, G., Tal, E. & Karaman, S. Multi-fidelity black-box optimization for time-optimal quadrotor maneuvers. Int. J. Robot. Res. 40 , 1352–1369 (2021).

Pham, H. & Pham, Q.-C. A new approach to time-optimal path parameterization based on reachability analysis. IEEE Trans. Robot. 34 , 645–659 (2018).

Song, Y., Romero, A., Müller, M., Koltun, V. & Scaramuzza, D. Reaching the limit in autonomous racing: optimal control versus reinforcement learning. Sci. Robot. (in the press).

Foehn, P. et al. Agilicious: open-source and open-hardware agile quadrotor for vision-based flight. Sci. Robot. 7 , eabl6259 (2022).

Jones, E. S. & Soatto, S. Visual-inertial navigation, mapping and localization: a scalable real-time causal approach. Int. J. Robot. Res. 30 , 407–430 (2011).

Download references

Acknowledgements

The authors thank A. Vanover, T. Bitmatta and M. Schaepper for accepting to race against Swift. The authors also thank C. Pfeiffer, T. Längle and A. Barden for their contributions to the organization of the race events and the drone hardware design. This work was supported by Intel’s Embodied AI Lab, the Swiss National Science Foundation (SNSF) through the National Centre of Competence in Research (NCCR) Robotics and the European Research Council (ERC) under grant agreement 864042 (AGILEFLIGHT).

Author information

Authors and affiliations.

Robotics and Perception Group, University of Zurich, Zürich, Switzerland

Elia Kaufmann, Leonard Bauersfeld, Antonio Loquercio & Davide Scaramuzza

Intel Labs, Munich, Germany

Matthias Müller

Intel Labs, Jackson, WY, USA

Vladlen Koltun

You can also search for this author in PubMed   Google Scholar

Contributions

E.K. formulated the main ideas, implemented the system, performed the experiments and data analysis and wrote the paper. L.B. contributed to the main ideas, the experiments, data analysis, paper writing and designed the graphical illustrations. A.L. formulated the main ideas and contributed to the experimental design, data analysis and paper writing. M.M. contributed to the experimental design, data analysis and paper writing. V.K. contributed to the main ideas, the experimental design, the analysis of experiments and paper writing. D.S. contributed to the main ideas, experimental design, analysis of experiments, paper writing and provided funding.

Corresponding author

Correspondence to Elia Kaufmann .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Peer review

Peer review information.

Nature thanks Sunggoo Jung, Karime Pereida and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables

Extended data fig. 1 residual models..

a , Visualization of the residual observation model and the residual dynamics model identified from real-world data. Black curves depict the residual observed in the real world and coloured lines show 100 sampled realizations of the residual observation model. Each plot depicts an entire race, that is, three laps. b , Predicted residual observation for a simulated rollout. Blue, ground-truth position provided by the simulator; orange, perturbed position generated by the Gaussian process residual.

Extended Data Fig. 2 Multi-iteration fine-tuning.

Rollout comparison after fine-tuning the policy for one iteration (blue) and two iterations (orange).

Supplementary information

Peer review file, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and Permissions

About this article

Cite this article.

Kaufmann, E., Bauersfeld, L., Loquercio, A. et al. Champion-level drone racing using deep reinforcement learning. Nature 620 , 982–987 (2023). https://doi.org/10.1038/s41586-023-06419-4

Download citation

Received : 05 January 2023

Accepted : 10 July 2023

Published : 30 August 2023

Issue Date : 31 August 2023

DOI : https://doi.org/10.1038/s41586-023-06419-4

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Toggle navigation

• Earlier Conferences
• Conference Help
• Create an Account
• Reset Password
• Merge Profiles
• Privacy Policy
• My Stuff/Registrations
• Edit Profile
• Change Password
• Create New Profile

NeurIPS 2023

Conference Dates: (In person) December 10 - December 16 

Homepage: https://neurips.cc/Conferences/2023/

Call For Papers 

Abstract submission deadline:  May 11, 2023 

Full paper submission (all authors must have an OpenReview profile when submitting) deadline: May 17, 2023 

Supplemental material submission deadline: May 24, 2023 

Camera-ready, poster, and video submission: to be announced

Submit at:   https://openreview.net/group?id=NeurIPS.cc/2023/Conference

The site will start accepting submissions on April 19, 2023 . 

Subscribe to these and other dates on the 2023 dates page .

The Thirty-Seventh Annual Conference on Neural Information Processing Systems (NeurIPS 2023) is an interdisciplinary conference that brings together researchers in machine learning, neuroscience, statistics, optimization, computer vision, natural language processing, life sciences, natural sciences, social sciences, and other adjacent fields. We invite submissions presenting new and original research on topics including but not limited to the following:

• Applications (e.g., vision, language, speech and audio)
• Deep learning (e.g., architectures, generative models, optimization for deep networks)
• Evaluation (e.g., methodology, meta studies, replicability and validity)
• General machine learning (supervised, unsupervised, online, active, etc.)
• Infrastructure (e.g., libraries, improved implementation and scalability, distributed solutions)
• Machine learning for sciences (e.g. climate, health, life sciences, physics, social sciences)
• Neuroscience and cognitive science (e.g., neural coding, brain-computer interfaces)
• Optimization (e.g., convex and non-convex, stochastic, robust)
• Probabilistic methods (e.g., variational inference, causal inference, Gaussian processes)
• Reinforcement learning (e.g., decision and control, planning, hierarchical RL, robotics)
• Social and economic aspects of machine learning (e.g., fairness, interpretability, human-AI interaction, privacy, safety, strategic behavior)
• Theory (e.g., control theory, learning theory, algorithmic game theory)

Machine learning is a rapidly evolving field, and so we welcome interdisciplinary submissions that do not fit neatly into existing categories.

Authors are asked to confirm that their submissions accord with the NeurIPS code of conduct .

Formatting instructions:  All submissions must be in PDF format. Submissions are limited to nine content pages, including all figures and tables; additional pages containing references are allowed. You must format your submission using the NeurIPS 2023 LaTeX style file which includes a “preprint” option for non-anonymous preprints posted online. The maximum file size for submissions is 50MB. Submissions that violate the NeurIPS style (e.g., by decreasing margins or font sizes) or page limits may be rejected without further review. If your submission is accepted, you will be allowed an additional content page for the camera-ready version. Papers may be rejected without consideration of their merits if they fail to meet the submission requirements, as described in this document. 

Use of Large Language Models (LLMs): We welcome authors to use any tool that is suitable for preparing high-quality papers and research. However, we ask authors to keep in mind two important criteria. First, we expect papers to fully describe their methodology, and any tool that is important to that methodology, including the use of LLMs, should be described also. For example, authors should mention tools (including LLMs) that were used for data processing or filtering, visualization, facilitating or running experiments, and proving theorems. It may also be advisable to describe the use of LLMs in implementing the method (if this corresponds to an important, original, or non-standard component of the approach). Second, authors are responsible for the entire content of the paper, including all text and figures, so while authors are welcome to use any tool they wish for writing the paper, they must ensure that all text is correct and original.

Double-blind reviewing:  All submissions must be anonymized and may not contain any identifying information that may violate the double-blind reviewing policy.  This policy applies to any supplementary or linked material as well, including code.  If you are including links to any external material, it is your responsibility to guarantee anonymous browsing.  Please do not include acknowledgements at submission time. If you need to cite one of your own papers, you should do so with adequate anonymization to preserve double-blind reviewing.  For instance, write “In the previous work of Smith et al. [1]…” rather than “In our previous work [1]...”). If you need to cite one of your own papers that is in submission to NeurIPS and not available as a non-anonymous preprint, then include a copy of the cited anonymized submission in the supplementary material and write “Anonymous et al. [1] concurrently show...”).

OpenReview: We are using OpenReview to manage submissions. The reviews and author responses will not be public initially (but may be made public later, see below). As in previous years, submissions under review will be visible only to their assigned program committee. We will not be soliciting comments from the general public during the reviewing process. Anyone who plans to submit a paper as an author or a co-author will need to create (or update) their OpenReview profile by the full paper submission deadline. Your OpenReview profile can be edited by logging in and clicking on your name in  https://openreview.net/ . This takes you to a URL "https://openreview.net/profile?id=~[Firstname]_[Lastname][n]" where the last part is your profile name, e.g., ~Wei_Zhang1. The OpenReview profiles must be up to date, with all publications by the authors, and their current affiliations. The easiest way to import publications is through DBLP but it is not required, see  FAQ . Submissions without updated OpenReview profiles will be desk rejected. The information entered in the profile is critical for ensuring that conflicts of interest  and reviewer matching are handled properly. Because of the rapid growth of NeurIPS, we request that all authors help with reviewing papers, if asked to do so.  We need everyone’s help in maintaining the high scientific quality of NeurIPS.  

Abstract Submission: There is a mandatory abstract submission deadline on May 11, 2023 01:00 PM PDT, six days before full paper submissions are due. While it will be possible to edit the title and abstract until the full paper submission deadline, submissions with “placeholder” abstracts that are rewritten for the full submission risk being removed without consideration. This includes titles and abstracts that either provide little or no semantic information (e.g., "We provide a new semi-supervised learning method.") or describe a substantively different claimed contribution.  The author list cannot be changed after the abstract deadline. After that, authors may be reordered, but any additions or removals must be justified in writing and approved on a case-by-case basis by the program chairs only in exceptional circumstances. 

Supplementary material:  Authors may submit up to 100MB of supplementary material, such as appendices, proofs, derivations, data, or source code; all supplementary materials must be in PDF or ZIP format. Supplementary material should be material created by the authors that directly supports the submission content. Like submissions, supplementary material must be anonymized. Looking at supplementary material is at the discretion of the reviewers. The deadline for supplementary material is one week after the main paper.

We encourage authors to upload their code and data as part of their supplementary material in order to help reviewers assess the quality of the work. Check the policy as well as code submission guidelines and templates for further details.

Ethics review:  Reviewers and ACs may flag submissions for ethics review . Flagged submissions will be sent to an ethics review committee for comments. Comments from ethics reviewers will be considered by the primary reviewers and AC as part of their deliberation. They will also be visible to authors, who will have an opportunity to respond.  Ethics reviewers do not have the authority to reject papers, but in extreme cases papers may be rejected by the program chairs on ethical grounds, regardless of scientific quality or contribution.   

Paper checklist: In order to improve the rigor and transparency of research submitted to and published at NeurIPS, authors are required to complete a paper checklist . The paper checklist is intended to help authors reflect on a wide variety of issues relating to responsible machine learning research, including reproducibility, transparency, research ethics, and societal impact. The checklist does not count towards the page limit and will be entered in OpenReview.

Preprints:  The existence of non-anonymous preprints (on arXiv or other online repositories, personal websites, social media) will not result in rejection. If you choose to use the NeurIPS style for the preprint version, you must use the “preprint” option rather than the “final” option. Reviewers will be instructed not to actively look for such preprints, but encountering them will not constitute a conflict of interest. Authors may submit anonymized work to NeurIPS that is already available as a preprint (e.g., on arXiv) without citing it. Note that public versions of the submission should not say "Under review at NeurIPS" or similar.

Dual submissions:  Submissions that are substantially similar to papers that the authors have previously published or submitted in parallel to other peer-reviewed venues with proceedings or journals may not be submitted to NeurIPS. Papers previously presented at workshops are permitted, so long as they did not appear in a conference proceedings (e.g., CVPRW proceedings), a journal or a book.  NeurIPS coordinates with other conferences to identify dual submissions.  The NeurIPS policy on dual submissions applies for the entire duration of the reviewing process.  Slicing contributions too thinly is discouraged.  The reviewing process will treat any other submission by an overlapping set of authors as prior work. If publishing one would render the other too incremental, both may be rejected.

Author responses:  Authors will have one week to view and respond to initial reviews. Author responses may not contain any identifying information that may violate the double-blind reviewing policy. Authors may not submit revisions of their paper or supplemental material, but may post their responses as a discussion in OpenReview. This is to reduce the burden on authors to have to revise their paper in a rush during the short rebuttal period.

After the initial response period, authors will be able to respond to any further reviewer/AC questions and comments by posting on the submission’s forum page. The program chairs reserve the right to solicit additional reviews after the initial author response period.  These reviews will become visible to the authors as they are added to OpenReview, and authors will have a chance to respond to them.

After the notification deadline, accepted and opted-in rejected papers will be made public and open for non-anonymous public commenting. Their anonymous reviews, meta-reviews, author responses and reviewer responses will also be made public. Authors of rejected papers will have two weeks after the notification deadline to opt in to make their deanonymized rejected papers public in OpenReview.  These papers are not counted as NeurIPS publications and will be shown as rejected in OpenReview.

Publication of accepted submissions:  Reviews, meta-reviews, and any discussion with the authors will be made public for accepted papers (but reviewer, area chair, and senior area chair identities will remain anonymous). Camera-ready papers will be due in advance of the conference. All camera-ready papers must include a funding disclosure . We strongly encourage accompanying code and data to be submitted with accepted papers when appropriate, as per the code submission policy . Authors will be allowed to make minor changes for a short period of time after the conference.

Contemporaneous Work: For the purpose of the reviewing process, papers that appeared online within two months of a submission will generally be considered "contemporaneous" in the sense that the submission will not be rejected on the basis of the comparison to contemporaneous work. Authors are still expected to cite and discuss contemporaneous work and perform empirical comparisons to the degree feasible. Any paper that influenced the submission is considered prior work and must be cited and discussed as such. Submissions that are very similar to contemporaneous work will undergo additional scrutiny to prevent cases of plagiarism and missing credit to prior work.

Plagiarism is prohibited by the NeurIPS Code of Conduct .

Other Tracks: Similarly to earlier years, we will host multiple tracks, such as Datasets and Benchmarks ,  Competitions , tutorials as well as workshops, in addition to the main track for which this call for papers is intended. See the conference homepage for updates and calls for participation in these tracks. 

Experiments: As in past years, the program chairs will be measuring the quality and effectiveness of the review process via randomized controlled experiments. All experiments are independently reviewed and approved by an Institutional Review Board (IRB).

Financial Aid : Each paper may designate up to one (1) NeurIPS.cc account email address of a corresponding student author who confirms that they would need the support to attend the conference, and agrees to volunteer if they get selected. To be considered for Financial the student will also need to fill out the Financial Aid application when it becomes available.

Amir Globerson, Moritz Hardt, Sergey Levine, Kate Saenko NeurIPS 2023 Program Chairs

Tristan Naumann and Alice Oh NeurIPS 2023 General Chairs