Download
about how to do research n.
Skip this Video
Loading SlideShow in 5 Seconds..
About how to do research PowerPoint Presentation
Download Presentation
About how to do research

About how to do research

120 Views Download Presentation
Download Presentation

About how to do research

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

  1. About how to do research Some tips for Graduate and PhD students (collected and adapted by A Florea)

  2. Content • Choosing a Research Project • Standard Pitfalls for Graduate Students • How to do it • Guide to reading • Progress • About your research report

  3. 1. Choosing a Research Project • You must be enthusiastic about it • Solving the problems it entails must be worthy of a MSc. or Ph.D. • It must be within sight of the state of the art, i.e. it must be ‘doable' in the time frame

  4. 2. Standard Pitfalls for Graduate Students • Solving the World – do not try to solve all problems in the world; set yourself a high but well focused goal • Manna from Heaven – do not stay in expectation of the "big" idea coming from heaven (as it will not usually come); be proactive, search for inspiration and ideas, read, talk, try to find examples and use cases • Computer Bum – do not rush to implement a program; think first about what you want to achieve, how, what method; make a description of the approach, define necessary algorithms, do not go to implementation unless you have at least a software architecture and a functional specification

  5. Standard Pitfalls for Graduate Students • Micawberism Wilkins Micawber, a character from Dickens’s novel David Copperfield , was always imagining himself as already achieving great things, then he made detailed plans of what to do once these great things were achieved Do not think you will fail; imagine that you have already achieved what you aim for; how does it look, how can it be used? (I am purposefully using present and not future tense, you should do the same)

  6. Standard Pitfalls for Graduate Students • Ivory Tower – do not shut yourself in an ivory tower to do your research (as this will not help your progress); talk to people, with colleagues, professors, people with similar interests but also to people who know nothing about your research • Misunderstood Genius – do not think that what you want to achieve or what you have done can not be understood by anybody; you should be able to explain to any person what you are working on, what it is good for, what your contribution is • Ambition Paralysis – do not wait to have the "great" idea; you may have it but, just in case you do not, start from small ideas, pick an example, try to solve it, and then try to apply the solution to some other examples. Sooner or later the "big" idea will come. “Work is 1% inspiration plus 99% transpiration”, Albert Einstein

  7. 3. How to do it • Define the problem / question • Form a Hypothesis/Hypotheses • Locate Resources/Gather Information & Materials • Plan the Research/Developing Methods • Develop a model & algorithm / approach • Set up/find a case and see if it fits

  8. How to do it • Design an architecture of your system • Write functional specifications • Implement it • Collect data and more cases • Test, interpret results • Develop an application(s) and think how to go further • Some steps may be re-iterated (due to feed-back) but do not skip steps

  9. How to do it During all steps, do not forget/hesitate (or whatever else might prevent you) to: • Communicate results • Do not wait to start writing; write about what you have done or what you aim to achieve – this will help you clarify the ideas and will allow you to obtain early feed-back • Talk to people – people knowledgeable in the domain but also other people – can you explain your idea to people who know nothing about research?

  10. 4. Guide to reading 3 levels of reading • Outer Circle • Middle Circle • Inner Circle

  11. Guide to reading • Outer Circle – articles about the surrounding problem (e.g, intelligent agents negotiation) • read an introductory material - all • read 1-2 surveys on negotiation – all and in depth • read some papers to become familiar with the research on intelligent agents negotiation: read the abstract, 50% of intro, conclusions + titles of sections

  12. Guide to reading • Middle Circle – articles about the particular type of problem you have in mind (e.g., heuristic negotiation) • read as for Outer Circle + all Intro + 1-2 core Sections • mark candidates for Inner circle • collect further readings from references

  13. Guide to reading • Inner Circle – articles about the specific approach you consider (e.g, heuristic negotiation strategies for multiple, concurrent bi-lateral encounters ) • read the entire article, then read it again • select key articles for your work and read them again (a 3rd time) • search applications based on the approach (you should do this for the middle circle also)

  14. Guide to reading For every article you have read (regardless of the circle) do: • note title, authors, journal, pages, etc., so as to able to refer it afterwards • make some notes about the content of every article (usually this will help you both to remember and to understand better) • length of notes should depend on the circle level

  15. 5. Progress • Try to keep your own progress report (your teacher may ask for one) • Achievements over the past xmonths • Goals for the next x months • Name 3 or 4 key papers that you have read in the past x months • (choose the x value that best fits you) • Collect other people feedback on your progress

  16. Bibliography about HtdR • http://www.cs.indiana.edu/mit.research.how.to/mit.research.how.to.html • http://www.ifs.tuwien.ac.at/~silvia/research-tips/ • http://depts.washington.edu/rural/RURAL/design/scimethod.html • One more tip: do not take for granted everything you read about HtdR (including these slides ) • Think about it, try it = question everything (this is what a researcher does), and find your own way to do it

  17. 6. About your research report • By the end of first semester you have to write a research report Requirements for the RR of Master AI students • Min 15 pages • Written in English • Use your own words, even when describing other's work or quote (X says that "…")

  18. Rememer what you have to cover in your RR • Define the problem / question • Form a Hypothesis/Hypotheses • Locate Resources/Gather Information & Materials • Plan the Research/Developing Methods • Develop a model & algorithm / approach • Set up/find a case and see if it fits • Design an architecture of your system • Write functional specifications • Implement it • Collect data and more cases • Test, interpret results • Develop an application(s) and think how to go further

  19. About your research report • The RR that you will write by the end of the year will contain the description of all steps (1-12) listed in Section 3 (slides 6 and 7) • The RR for semester I must contain at least the description of: • 1,2 (problem, hypothesis) • 3 – presentation and synthesis of what you have read (state of the art) • 4, 5 – first version (plan, model, algorithm) • 6 (use case) • 7 – first version (architecture) • Ideally, it should include 8 (specs) – first version, and a demonstrator (a small program to demonstrate some of your ideas)

  20. About your research report • The RR must describe clearly your contribution(s) • Your main contribution(s) may be at different levels (one or several of, although not limited to): • a new/improved method • new/improved algorithm(s) • a new/improved system architecture • the application of an algorithm / method to solving a different problem from the ones existing in the literature • a new implementation with improved functionalities, enhanced user interface, etc. • a new application (even if based on existing algorithms, techniques) • creative synthesis of what you have read, new classification dimensions and future trends

  21. About your research report When presenting your contribution(s) you should mention: • Why you did it • How it differs from existing approaches (e.g, better, faster, more test cases, more problems to be solved, a different application, etc.) highlighting merits • Existing technologies and your foreseen one • Shortcomings, possible ways of removing them • How you will continue

  22. Bibliography http://www.ifs.tuwien.ac.at/~silvia/research-tips/Writing%20a%20paper.pdf - I do not support everything it is written here but there are good tips E. Robert Schulman, How to Write a Scientific Paper, Annals of Improbable Research, Vol. 2, No. 5, pg. 8. http://members.verizon.net/~vze3fs8i/air/airpaper.html - Not (very) serious but amusing