This is the repository for the collection of instrunctions about how to research.
- Choose the right lab
- Learn to think like a scientist.
- Look for questions, not subjects.
- Mentoring is as important as research.
- Reach out with confidence.
- Get the most out of the experience
- Start your research with reading, and keep on reading.
- Set specific goals for yourself and let your mentors know.
- Research takes time.
- Find a healthy balance.
- Find financial support.
- Always think about the big picture.
- Read broadly.
- Work on two things.
- Be ambitious.
- Collaborate.
- Be proactive.
- Write a blog.
- Keep a source of positive energy.
- Play to your strengths.
- Intern or visit a university.
- Play the long game.
Advice - A repository of links with advice related to grad school applications, research, phd etc related to (usually but not always) CS, NLP, and ML.
Researchers spend a great deal of time reading research papers. However, this skill is rarely taught, leading to much wasted effort. This article outlines a practical and efficient three-pass method for reading research papers. I also describe how to use this method to do a literature survey.
Comment: Highly Recommneded. Practical Formula with Annotated Examples
- General Tips
- Abstract
- General and Specific Background (~1 sentence each). Introduce the area of science that you will be speaking about and the state of knowledge in that area. Start broad in the general background, then narrow in on the relevant topic that will be pursued in the paper. If you use jargon, be sure to very briefly define it.
- Knowledge Gap (~1 sentence). Now that you’ve stated what is already known, state what is not known. What specific question is your work attempting to answer?
- “Here we show…” (~1 sentence). State your general experimental approach and the answer to the question which you just posed in the “Knowledge Gap” section.
- Experimental Approach & Results (~1-3 sentences). Provide a high-level description of your most important methods and results. How did you get to the conclusion that you stated in the “Here we show…” section?
- Implications (~1 sentence). Describe how your findings influence our understanding of the relevant field and/or their implications for future studies.
- Introduction
- General Background. Introduce the general area of science in which your project takes place, highlighting the status of our understanding of that system.
- Specific Background. Narrow down to the sub-area that your paper will be addressing, and again highlight the extent of our understanding in this sub-area.
- Knowledge Gap. After discussing what we know, articulate what we do not know, specifically focusing on the question that has motivated your work.
- “Here we show…” Very briefly summarize your methods and findings.
- Methods
- State the reasons for choosing your methodology
- Use subheadings to organize content
- Provide minimal essential detail
- Avoid “we did…” or “the authors did…”
- Discussion
- Tell how your paper is special
- Relate your results to existing results
- Tell how your study’s limitations leave open the big questions
- Results
- Create a logical narrative, organized into subsections
- Each Results paragraph has rationale, data, and a transition
- Show minimal essential data
- Use appropriate style
Springer Nature Author tutorials: Writing a journal manuscript
By the end of the tutorial you should know on how to:
- Prepare prior to starting your research
- Structure your manuscript and what to include in each section
- Get the most out of your tables and figures so that they clearly represent your most important results.
- Background (2-3 sentences): Provide just enough background to “set the scene” – the bare minimum necessary to make what follows understandable.
- Question (1 sentence): What was the goal of your study? State your hypothesis or question clearly and succinctly.
- Results (4-5 sentences): I generally aim to sum up each figure or section of the paper in one sentence of the abstract.
- Conclusions (1-2 sentences): Sum up very quickly why your results are important by tying them back in to the issue you mentioned in the introductory sentence.
- Do Summarize Your Results and Outline Their Interpretation in Light of the Known Literature
- Do Explain the Importance of Your Results
- Do Acknowledge the Shortcomings of the Study
- Do Discuss Any Future Directions
- Don’t Reiterate Your Results
- Don’t Over-Interpret Your Findings
- Don’t Introduce a New Piece of Data
- Don’t Use Too Much Jargon
Twelve tips for students who wish to write and publish-University of Exeter Medical School-Rajaei K Sharma & Harriet L Ogle
(1) find your why; (2) play to your strengths and be realistic; (3) be well read; (4) revisit missed opportunities; (5) talk to the doctors around you; (6) broaden your horizons; (7) get to grips with the submission process early; (8) pay attention to the details; (9) remember that submission is not the end; (10) the process can’t be rushed; (11) consider the alternative paths to presenting research; (12) start writing.
This article presents the main challenges of academic writing and publication in scientific journals. It reveals some of the most common mistakes in the process of manuscript submission and review, and offers some possible solutions.
8 sequential steps to write a first rough draft of a research paper from start to finish (relatively quick and easy)
How to Write a Research Paper-Department of Computer Science-University of Illinois at Chicago-Natalie Parde
This article provides a step-by-step guide to help you turn your high-quality data into a high-quality manuscript for publication in a scientific journal. It covers all aspects of the writing process, including: choosing a journal to which to submit your paper, writing each section, formulating your “story,” making figures, soliciting constructive criticism, and navigating the review process.
Make it easy to say 'yes'. The first rule to get your research published is to make life easy for the editor and the reviewers. The task for an author is ultimately a narrative task - to succinctly and engagingly tell the story of what you did, why you did it, what you found, and why we should care - and making it clear why the paper is within the scope of the journal. The more carefully and meticulously a paper is prepared - including spelling, punctuation, grammar, structure, completeness and internal consistency - the more likely it is to be published and the quicker it will appear in print.
Formal literature reviews are a critical appraisal of a subject and are not only an academic requirement but essential when planning a research project and for placing research findings into context. Understanding the landscape in which you are working will enable you to make a valuable contribution to your field. Writing a literature review requires a range of skills to gather, sort, evaluate and summarise peer-reviewed published data into a relevant and informative unbiased narrative. Digital access to research papers, academic texts, review articles, reference databases and public data sets are all sources of information that are available to enrich your review.
This paper discusses the question about how to write a literature review paper (LRP). It stresses the primary importance of adding value, rather than only providing an overview, and it then discusses some of the reasons for (or not) actually writing an LRP, including issues relating to the nature and scope of the paper. It also presents different types of LRPs, advises on reporting the methodology used for the selection of papers for review, and the structure of an LRP. An important conclusion is that the heterogeneity in LRPs is very large. This paper also presents some of the aspects that the authors feel are important structural and contextual considerations that help produce high-quality review papers.
How to write a great abstract for your academic manuscript, Social Sciences & Humanities Open, Elsevier, September 2020
Writing for Impact: How to Prepare a Journal Article, Medical and Scientific Publishing, Elsevier, August 2018
Useful Phrases - Chapter 19 - English for Writing Research Papers-version-1
"听JEBO主编Daniel Houser讲论文写作。关于摘要，他说他心中的标准摘要应该有五句，第一句是论文在文献里的定位，第二句是这篇文章做什么，第三句是为什么做，第四句是发现了什么，第五句是这些发现有什么意义。然后这五句其实各自对应文章的一部分，第一句是引言，第二句是研究设计，第三句是待检验假设或文献讨论，第四句是研究结果，第五句是结论。也就是说，摘要每一句话都是正文一个部分的总结。" （忘记出处了）
An IJDS paper will have four ingredients, data + methodology + decisions + implications, but the key is the synthesis of the ingredients, anchored to methodology. Here is one potential approach for such synthesis that will be familiar to many in our community:
Fired by a managerial/industrial decision-making motivation and potential/actual impact, the paper introduces aninnovative data sciencemethodology (model/algorithm/ approach), applies it to data (real-world and/or simulated) to illustrate its usefulness and behavior, and considers practical (e.g., computation, implementation) and ethical (e.g., societal, environmental) implications.
Each author may include a photo and biography in their article, although photos and biographies are often omitted from conference-related articles. Follow these tips for writing your biography.
- If you are an IEEE member, list your status and the corresponding year in parentheses immediately after your name; for example, (M'01-SM'05-F'15) denotes that you became an IEEE Member (M) in 2001, a Senior Member (SM) in 2005, and a Fellow (F) in 2015.
- The first paragraph may contain your place and/or date of birth, followed by your educational background. Degrees should be listed with the type of degree in what field, which institutions, city, state, and country, and the year the degree was earned. Your major field of study should be in the lower-case.
- The second paragraph lists military and work experience, with job titles capitalized. Include a location for your current job. List no more than three published books or articles. Include a list of current and previous research interests. In the second paragraph, use the appropriate pronoun and not your last name.
- The third paragraph begins with your title and last name (e.g., Dr. Smith, Prof. Chen) and lists any memberships in professional societies other than IEEE. List any awards and work for IEEE committees and publications.
- If you choose to include a photo, it should be a professional, high-quality headshot. Note that while photos are optional, if they are included, they will be published.
- Remember, biographies and photos cannot be edited or removed after article publication.
What makes an innovative materials study? Conceptual novelty is intrinsically difficult to quantify. There are common pitfalls that lead to rejection without review, leading to an editorial shorthand of archetype rejections. Here, we describe the most common reasons for rejection at Matter.
IEEE Author Education course: Tips for Responding to Reviewers' Comments (video)
- Session 1: How to write a good paper
- Session 2: How to write a good review
- Session 3: How to write a good rebuttal?
- Session 4: How are reviews used in the decision process?
- Video (Bilibili)
- Website (in Chinese)
- How to Properly Review AI Papers?-University of Rochester-Jiebo Luo (Slides)
- How to Get Your CVPR Paper Rejected?-University of California at Merced-Ming-Hsuan Yang (Slides)
- Chinese Summary
- Read the room.
- Be clear about your main message.
- Deliver your takeaway at the start.
- Make a plan.
- Storyboard your presentation.
- Be kind to your audience.
- Allow your audience to listen, not read.
- Use pictures to connect on a human level.
- Create visuals for the back row.
- Simplify and enlarge.
- Explain your figures.
- Write ‘sentence headline’ titles.
- Use a lot of slides if you want.