Presentation on theme: "How to Read a Research Paper"— Presentation transcript:
1How to Read a Research Paper Tamer ElsayedFaculty of Engineering Computer Science and Engineering Dept.May 16th 2013
2A Typical Researcher …Will likely spend hundreds of hours every year reading papers A “good” graduate student /researcher should read (in average) a paper a day!
3ReferencesS. Keshav, How to Read a Paper, ACM SIGCOMM Computer Communication Review, 2007Philip W. L. Fong, Reading a Computer Science Research Paper, SIGCSE 2009Amanda Stent, How to Read a Computer Science Research Paper, Technical Report.Mihai Pop, How to Read a Scientific Paper, a Presentation at University of Maryland, College Park.
5Why?You were asked to For a literature survey of a new field/problemBe up-to-date on current research in the fieldAllows you to replicate/extend the resultsProvides you with useful dataGives you “pre-digested” thoughtsTo decide whether (and where) to publishTeaches you how to writeReview for a conference or a class….
6Roadmap Types of (computing) research papers √ Why reading research papers?Types of (computing) research papersThree-pass approach for reading a paperDoing a literature reviewHow can I remember the papers I have read?
7Types of (Computing) Research Papers From venue, peer-review, and content perspectives …
8Venue? Most recent, “hot off the presses”, information Conference PapersMost recent, “hot off the presses”, informationTechnical ReportsExpands on the information in a conference paper.Journal Papers (or Articles)Expand and combine results from several conference papersBook ChaptersExpand a conference paper or journal articlePhD DissertationsFrequently revised into either journal papers or book chaptersPostersExtended abstracts, short, good for discussionsWorkshop PapersVery hot/new topics, work in-progress, ideas, preliminary results
9Peer-Review? Conference papers and journal papers are “peer-reviewed” have been examined by other computer scientists (3-5) prior to publication.“Double-blind” reviewTechnical reports are typically not peer-reviewed, but are still excellent sources of detailed information .Posters and workshop papers are peer-reviewed (less number of reviewers).MSc thesis or PhD dissertation?
10Content? Theoretical Engineering Empirical Survey Describe/prove a theoryDescribe new algorithmsTheoreticalDescribe an implementation of an algorithm, or part or all of a computer system or applicationEngineeringDescribe an experiment designed to test some hypothesisEmpiricalReview current results in a field of researchSurvey
11A Three-pass approach for reading a research paper Each pass accomplishes specific goals and builds upon the previous pass …
12Three-Pass Approach First pass Second pass Third pass Gives you a general idea about the paper.First passLet you grasp the paper’s content, but not its details.Second passHelps you understand the paper in depth.Third passThe key idea is that you should read the paper in up tothree passes, instead of starting at the beginning and plowing your way to the end.
131 The First Pass A quick scan to get a bird’s-eye view of the paper. Should take about 5-10 minutes.Decide whether you need to do any more passes.Carefully read the title, abstract, and introductionRead the section and sub-section headings, but ignore everything elseRead the conclusionGlance over the references, mentally ticking off the ones you’ve already read1
141 At the End of this Pass … More passes? You should be able to answer the five Cs:Category: What type of paper is this? A measurement paper? An analysis of an existing system? A description of a research prototype?Context: What is the problem (space)? Which other papers is it related to?Correctness: Do the assumptions appear to be valid?Contributions: What are the main contributions?Clarity: Is the paper well written?More passes?1
151 2 The Second Pass Should take up to 1 hour. Read with greater care, but ignore details such as proofs.Identify key points, or make comments in the margins.Look carefully at the figures, diagrams and other illustrations in the paper.Remember to mark relevant unread references for further readingthis is a good way to learn more about the background of the paper.12
16After this Pass …Sometimes, you won’t understand it even at the end of the second pass subject is new to you, with unfamiliar terminology and acronyms.proof or experimental technique that you don’t understand.poorly written: unsubstantiated assertions and numerous forward references.it could just be that it’s late at night and you’re tired!You can now choose to:set the paper aside, hoping you don’t need to understand the material to be successful in your career,return to the paper later, perhaps after reading background material orpersevere and go on to the third pass.12
171 2 3 The Third Pass To fully understand the paper (e.g., reviewing) Can take about 4-5 hours for beginners and about 1 hour for an experienced reader.Requires great attention to detail.Attempt to virtually re-implement the paper: making the same assumptions as the authors, re-create the work.You should identify and challenge every assumption in every statement.During this pass, you should also jot down ideas for future work.123
18At the End of this Pass …Should be able to reconstruct the entire structure of the paper from memory.Should be able to identify its strong and weak points.Should be able to pinpoint implicit assumptions, missing citations to relevant work, and potential issues with experimental or analytical techniques.123
20Doing a Literature Survey What is it?Requires you to read tens of papers, perhaps in an unfamiliar field.What papers should you read?3-pass approach to help.
21First StepUse an academic search engine and some well-chosen keywords to find 3 to 5 recent papers in the area.(e.g., Google Scholar, MS Academic Search, CiteSeer, or ACM digital library)Do one pass on each paper to get a sense of the work, then read their related work sections.You will find a summary of the recent workand perhaps, if you are lucky, a recent survey paper.If you can find such a survey, you are done.Read the survey, congratulating yourself on your good luck.1
22Second StepFind shared citations and repeated authors in the bibliography.These are the key papers and researchers in that area.Download the key papers and set them aside.Then go to the websites of the key researchers and see where they’ve published recently.That will help you identify the top conferences in that field12
23Third StepGo to the website for these top conferences and look through their recent proceedings.A quick scan will usually identify recent high-quality related work.These papers, along with the ones you set aside earlier, constitute the first version of your survey.Make two passes through these papers.If they all cite a key paper that you did not find earlier, obtain and read it, iterating as necessary.123
25Remembering Read-Papers Strongly recommend that you make an electronic file for your own bibliography.A BibTeX file is a good idea.After reading, if you think it’s worth remembering, write an entry for that paper in your bibliography file.You should note: authors’ names, paper title, venue, date of publication, and page numbers (if possible).Add a 2-3 sentence description of the paper in which you summarize for yourself the problem addressed by the paper, the solution proposed or results learned, and main contributions.
26Other ResourcesReading scientific papers (at Purdue)BibTeXMendeleyZotero – firefox extension reference managerComparison of reference manager software tools available
28Workshop next Semester … Stay Tuned How to Do ResearchResearch methodologyHow to read a research paperHow to review a paperHow to write a research paperHow to write a lit reviewHow to write theses/dissertationsHow to be a good graduate studentHow to give a talkHow to present a posterWorkshop next Semester … Stay Tuned
29Acknowledgement To my colleagues : Maram Hasanain Latifa AlMarri Rana Malhas