Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Research Skills I Learned in UIUC from Pi-Cheng Hsiu

2,641 views

Published on

Published in: Education, Technology
  • Be the first to comment

Research Skills I Learned in UIUC from Pi-Cheng Hsiu

  1. 1. Research Skills I learned in UIUC Department of Computer Science and Information Engineering, National Taiwan University Nov. 6th, 2007 Pi-Cheng Hsiu
  2. 2. Outline <ul><li>Pick a research area </li></ul><ul><li>Identify a research problem and solve it </li></ul><ul><li>Write a paper </li></ul><ul><li>Submit to a conference </li></ul><ul><li>Give a presentation </li></ul>05/28/10
  3. 3. Research Types <ul><li>Research is a re peat search process to find out the unknown. – Lui Sha </li></ul>05/28/10
  4. 4. Life of Ra (a Research Area) 05/28/10 – Indranil Gupta
  5. 5. Researcher Styles – Tarek Abdelzaher <ul><li>Salesman </li></ul><ul><ul><li>I have an idea to sell. What’s the market segment where it can make the most impact? </li></ul></ul><ul><li>Consultant </li></ul><ul><ul><li>Focus on problem P from company X. How do I solve P? </li></ul></ul><ul><li>Matchmaker </li></ul><ul><ul><li>Idea A in field X is a really good match to problem B in field Y. Apply A to B. </li></ul></ul><ul><li>Journalist </li></ul><ul><ul><li>Is this an interesting topic today? Create a new direction. </li></ul></ul>05/28/10
  6. 6. Key to Finding Your Research Area <ul><li>Match your interest/passion (theory vs. system) </li></ul><ul><ul><li>But don ’ t limit yourself within your comfortable zone </li></ul></ul><ul><li>Talk with other people and get feedback </li></ul><ul><ul><li>But have you own insights </li></ul></ul><ul><li>Read best papers on top conferences </li></ul><ul><ul><li>But do not always be a follower </li></ul></ul><ul><li>Study broadly </li></ul><ul><ul><li>But be sure you know what the field is really about </li></ul></ul>05/28/10
  7. 7. Outline <ul><li>Pick a research area </li></ul><ul><li>Identify a research problem and solve it </li></ul><ul><li>Write a paper </li></ul><ul><li>Submit to a conference </li></ul><ul><li>Give a presentation </li></ul>05/28/10
  8. 8. Keep in Mind <ul><li>It is difficult to say what is impossible, for the dream of yesterday is the hope of today and the reality of tomorrow.&quot; </li></ul><ul><li> – Robert H. Goddard </li></ul><ul><li>Don ’ t try to “ solve the world ” or “ boil the ocean ” . </li></ul>05/28/10
  9. 9. The Tough Part <ul><li>Finding a problem is 90% of the problem. – P. R. Kumar </li></ul><ul><ul><li>Open problems (e.g., P = NP?) </li></ul></ul><ul><ul><ul><li>What is solvable? </li></ul></ul></ul><ul><ul><li>Mature areas </li></ul></ul><ul><ul><ul><li>What is unknown? </li></ul></ul></ul><ul><ul><li>Emerging areas </li></ul></ul><ul><ul><ul><li>What is promising? </li></ul></ul></ul>05/28/10
  10. 10. Identifying the Problem <ul><li>Identifying problems is more important than finding solutions </li></ul><ul><ul><ul><li>– Y. Y. Zhou </li></ul></ul></ul><ul><li>Define the boundaries of your problem carefully </li></ul><ul><ul><li>Nobody will be impressed if you set the bar too low and jump over it. </li></ul></ul><ul><ul><li>Nobody will be impressed if you set the bar too high and don ’ t jump over it. (Dave Redell) </li></ul></ul>05/28/10
  11. 11. Sharpen Your Tools <ul><li>Take a solid set of foundational courses </li></ul><ul><ul><li>Math courses </li></ul></ul><ul><ul><ul><li>Graph Theory, Combinatorics, Algebra, Probability Theory, Stochastic Processes, Algorithms, etc. </li></ul></ul></ul><ul><ul><li>Computer Science courses </li></ul></ul><ul><ul><ul><li>Programming Language, Operating Systems, Computer Architecture, Computer Networks, Compiler, etc. </li></ul></ul></ul><ul><li>Technical English Writing </li></ul>05/28/10
  12. 12. Core of Learning (Edgar Dale) <ul><li>After 2 weeks, we tend to remember: </li></ul><ul><ul><li>10% of what we </li></ul></ul><ul><ul><li>20% of what we </li></ul></ul><ul><ul><li>30% of what we </li></ul></ul><ul><ul><li>50% of what we </li></ul></ul><ul><ul><li>70% of what we </li></ul></ul><ul><ul><li>90% of what we </li></ul></ul><ul><li>The more energy that I put into a subject, the more I can remember. </li></ul><ul><li> – Lui Sha </li></ul>05/28/10 read hear see (pictures) hear and see say say and do
  13. 13. Outline <ul><li>Pick a research area </li></ul><ul><li>Identify a research problem and solve it </li></ul><ul><li>Write a paper </li></ul><ul><li>Submit to a conference </li></ul><ul><li>Give a presentation </li></ul>05/28/10
  14. 14. Keep in Mind <ul><li>You could write a great paper even with research that is just ok. You can also write a bad paper despite good research. </li></ul><ul><li>– A. Bhatele, P. Jetley, and A. Becker in UIUC </li></ul>05/28/10
  15. 15. Core Document <ul><li>Mary Shaw, Writing Good Software Engineering Research Papers . In Proc. of the 25th International Conference on Software Engineering (ICSE 2003), p726-736. </li></ul><ul><li>Answer Mary Shaw ’ s questions first </li></ul><ul><ul><li>Helps you refine ideas </li></ul></ul><ul><ul><li>Helps you communicate ideas </li></ul></ul><ul><ul><li>Helps you organize paper writing </li></ul></ul>05/28/10
  16. 16. Key to Writing <ul><li>Find a model paper </li></ul><ul><li>Follow the three Bs </li></ul><ul><ul><li>Brevity </li></ul></ul><ul><ul><ul><li>Say it simply. </li></ul></ul></ul><ul><ul><li>Balance </li></ul></ul><ul><ul><ul><li>Balance the formal with the informal. </li></ul></ul></ul><ul><ul><li>Benefit </li></ul></ul><ul><ul><ul><li>Write for the benefit of your audience. </li></ul></ul></ul>05/28/10 – Tanya L. Crenshaw
  17. 17. Paper Organization <ul><li>Abstract: 100-150 words </li></ul><ul><li>Introduction: be brief! </li></ul><ul><ul><li>Related work </li></ul></ul><ul><li>Body of paper </li></ul><ul><ul><li>Problem (sufficient motivation & at least one scenario) </li></ul></ul><ul><ul><li>Approach, architecture </li></ul></ul><ul><ul><li>Experiment results </li></ul></ul><ul><li>Conclusion and future work </li></ul>05/28/10
  18. 18. Abstract (general rules) <ul><li>No references </li></ul><ul><li>Avoid use of “in this paper” </li></ul><ul><li>Avoid general motivation in the abstract </li></ul><ul><li>Highlight not just the problem, but also the principal results </li></ul><ul><li>Have terms that identify your work since abstract will be used by search engines </li></ul><ul><ul><li>The name of any protocol or system developed </li></ul></ul><ul><ul><li>The general area </li></ul></ul><ul><li>Avoid questions and math </li></ul>05/28/10
  19. 19. Introduction (recipe) <ul><li>Paragraph 1: Motivation (high level) </li></ul><ul><ul><li>Set the larger context </li></ul></ul><ul><li>Paragraph 2: Explain problem </li></ul><ul><ul><li>Establish specific context and background </li></ul></ul><ul><li>Paragraph 3: Compare/contrast related work </li></ul><ul><ul><li>What’s the general approach taken? </li></ul></ul><ul><ul><li>Show respect for peer-reviewed results </li></ul></ul><ul><li>Paragraph 4: Sell your work (**) </li></ul><ul><ul><li>Put your theme sentence here </li></ul></ul><ul><ul><li>Summarize your contributions </li></ul></ul><ul><ul><li>Why are the specific results significant? </li></ul></ul><ul><li>Paragraph 5: Outline the paper </li></ul><ul><ul><li>Avoid redundant phrasing </li></ul></ul>05/28/10
  20. 20. Body of Paper (tips) <ul><li>Avoid jargon </li></ul><ul><ul><li>Define before using unless well known </li></ul></ul><ul><li>Follow common terms </li></ul><ul><ul><li>Don’t use a common term for your own definition </li></ul></ul><ul><ul><li>Don’t use your own term for a common definition </li></ul></ul><ul><li>Balance formal and informal </li></ul><ul><ul><li>Ã x (A(x)) > Ã y (A(y)) vs. “x is better than y” </li></ul></ul><ul><li>Prefer simple to complicated </li></ul><ul><ul><li>“ now” versus “at this point in time” </li></ul></ul><ul><li>Obscure and pretentious use of language </li></ul><ul><ul><li>Don’t show off. Clarity is most important. </li></ul></ul>05/28/10
  21. 21. A Good Way to Learning How to Write <ul><li>Review papers and then learn what you think good and remember what you think bad. </li></ul><ul><li>– P.-C. Hsiu </li></ul>05/28/10
  22. 22. Outline <ul><li>Pick a research area </li></ul><ul><li>Identify a research problem and solve it </li></ul><ul><li>Write a paper </li></ul><ul><li>Submit to a conference </li></ul><ul><li>Give a presentation </li></ul>05/28/10
  23. 23. Prior Review (writers’ workshop) <ul><li>Designed for students to get early feedback from their peers </li></ul><ul><li>Send the draft to the other members </li></ul><ul><li>Meet to discuss </li></ul><ul><ul><li>Members discuss what they didn’t understand and didn’t like </li></ul></ul><ul><ul><li>The author is present but remains silent </li></ul></ul>05/28/10
  24. 24. Where to Submit to <ul><li>Good ones top ones, but reasonable ones </li></ul><ul><ul><li>Don’t submit just for kicks. Spoils your reputation </li></ul></ul><ul><ul><li>Don’t submit to a lower conference if you feel the paper can do well at a better conference </li></ul></ul><ul><ul><li>Ask your advisor </li></ul></ul><ul><ul><li>Go to conferences, make friends and contacts, and ask your friends about “reputation” of a conference </li></ul></ul><ul><li>Submit to journals! Takes a long time, but is very important! </li></ul>05/28/10
  25. 25. Outline <ul><li>Pick a research area </li></ul><ul><li>Identify a research problem and solve it </li></ul><ul><li>Write a paper </li></ul><ul><li>Submit to a conference </li></ul><ul><li>Give a presentation </li></ul>05/28/10
  26. 26. Sha ’ s I 3 Model <ul><li>An ideal presentation is one that is </li></ul><ul><ul><li>Informative </li></ul></ul><ul><ul><li>Interesting </li></ul></ul><ul><ul><li>Insightful </li></ul></ul>05/28/10
  27. 27. Being Informative <ul><li>Give new knowledge (audience-oriented) </li></ul><ul><li>“ New” is relatively to your audience . </li></ul><ul><ul><li>what they already know? </li></ul></ul><ul><ul><li>what they should know after your presentation? </li></ul></ul><ul><ul><li>what are the steps in-between? </li></ul></ul>05/28/10
  28. 28. Being Interesting <ul><li>Unexpected, counter intuitive, and difficult to believe </li></ul><ul><ul><li>Seemingly unimportant fact that actually holds the key. </li></ul></ul><ul><ul><li>Seemingly true but it is in fact false. </li></ul></ul><ul><ul><li>A “difficult” problem is solved with ease and elegance. </li></ul></ul>05/28/10
  29. 29. Being Insightful <ul><li>Impart a deeper understanding </li></ul><ul><ul><li>Explain a seemingly complex and confusing problem in a way that is easy to understand. </li></ul></ul><ul><ul><li>Unearth hidden/unstated assumptions and quickly put an argument to rest. </li></ul></ul><ul><ul><li>Show things in new angles, new lights and new forms and gain new understandings. </li></ul></ul><ul><ul><li>Demonstrate subtle but important connections/inter-dependencies between seemingly unrelated subjects. </li></ul></ul>05/28/10
  30. 30. An Example <ul><li>Using simplicity to control complexity </li></ul><ul><ul><li>Bubble sort: robust but slow </li></ul></ul><ul><ul><li>Quick sort: vulnerable but fast </li></ul></ul><ul><ul><li>How to have both? </li></ul></ul>05/28/10
  31. 31. Is There a Perfect Agenda? <ul><li>Generalized rules always have exceptions . </li></ul>05/28/10 Made by Mu Sun
  32. 32. Truths <ul><li>A rolling stone gathers no moss </li></ul><ul><li>Luck favors the one who is prepared </li></ul><ul><li>Practice makes perfect </li></ul>05/28/10
  33. 33. Conclusion <ul><li>The entrance to graduate school marks a critical phase of transition for most graduate students from absorbing knowledge to creating knowledge </li></ul><ul><ul><li>– Lui Sha </li></ul></ul><ul><li>Basically, a PhD degree means you are equipped to do all the following </li></ul><ul><ul><li>1. Pick an arbitrary area (not necessarily in CS) </li></ul></ul><ul><ul><li>2. Understand it yourself </li></ul></ul><ul><ul><li>3. Identify problems in it </li></ul></ul><ul><ul><li>4. Solve these problems in an innovative manner </li></ul></ul><ul><ul><li>– Indranil Gupta </li></ul></ul>05/28/10
  34. 34. Reference <ul><li>https://agora.cs.uiuc.edu/display/cs598lrs/Home </li></ul>05/28/10
  35. 35. Vote <ul><li>Q1: If you had taken this course, do you suggest it be consider a regular course? </li></ul><ul><ul><li>Yes vs. No = 41: 0 </li></ul></ul><ul><li>Q2: Do you want our department to have a similar course? </li></ul><ul><ul><li>Yes vs. No = 41: 0 </li></ul></ul>05/28/10

×