Raskar Thesis Guidance 2011


Published on

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • Six ways of coming up with new ideas based on an idea ‘X’. Ramesh Raskar Associate Professor MIT Media Lab http://raskar.info http://cameraculture.info http://raskar.info http://cameraculture.info
  • http://www.slideshare.net/cameraculture/raskar-ideahexagonapr2010
  • X up: Airbags for car, for helicopter
  • X up: Airbags for car, for helicopter
  • http://en.wikipedia.org/wiki/George_H._Heilmeier#Heilmeier.27s_Catechism
  • Five on Five = If more than five teams in the world are doing the same research, don’t do it. = If you disappear for five years, will someone do it anyway? Then your idea is not that great anyway. = Can you explain your work in five sentences to your grandma how it will impact human life? = If you can explain the idea in five minutes to a student and disappear for five years, will s/he be able to do it on her/his own without additional input from you/without iterations .. It is too obvious and lacks depth .. Don’t do it. = Strive to work on ideas that may require five+ disciplines .. Today’s research is highly team-driven and more diverse the required team composition, more fun you will have and also indicates a natural barrier to entry for others satisfying condition 1 and 2 Much like the food pyramid, five servings are the goal and will make you stronger .. But ok if your research project does not satisfy all five conditions
  • Raskar Thesis Guidance 2011

    1. 1. Preparing for a Masters Thesis Ramesh Raskar MIT Media Lab
    2. 2. <ul><li>Questions before you start the project </li></ul><ul><li>How to come up with ideas </li></ul><ul><li>How to write a paper </li></ul><ul><li>How to decide if the idea is worth pursuing </li></ul><ul><li>What makes a great ML thesis </li></ul><ul><li>Happy to meet </li></ul><ul><ul><li>Help you towards a fantastic + manageable thesis </li></ul></ul>Ramesh Raskar, http://raskar.info
    3. 3. <ul><li>An original piece of work </li></ul><ul><ul><li>A good thesis puts forth an original hypothesis/method/design/art piece with appropriate testing/verification/critique </li></ul></ul>What Makes a Good ML Thesis? <ul><li>Written well </li></ul><ul><ul><li>Grammar, appropriate style, organization </li></ul></ul><ul><li>Comprehensive </li></ul><ul><ul><li>A good thesis has sufficient information to allow a person of ordinary skill in the art to replicate the results </li></ul></ul><ul><ul><li>A good thesis has a complete set of references </li></ul></ul><ul><li>Accurate </li></ul><ul><ul><li>Do not blow hot air!! Every sentence in the thesis must be correct!!! Do not exaggerate!!! </li></ul></ul>Slide by Hugh Herr
    4. 4. Research .. <ul><li>http://raskar.info </li></ul><ul><ul><li>How to come up w ideas: Idea Hexagon </li></ul></ul><ul><ul><li>How to write a paper </li></ul></ul><ul><ul><li>How to give a talk </li></ul></ul><ul><ul><li>Open research problems </li></ul></ul><ul><ul><li>How to decide merit of a project </li></ul></ul><ul><ul><li>How to attend a conference, brainstorm </li></ul></ul><ul><li>Tips </li></ul><ul><ul><li>Get on Seminar/Talks mailing lists worldwide </li></ul></ul><ul><ul><li>http://www.cs.virginia.edu/~robins/YouAndYourResearch.html </li></ul></ul><ul><ul><li>Why do so few scientists make significant contributions and so many are forgotten in the long run? </li></ul></ul><ul><ul><li>Highly recommended Hamming talk at Bell Labs </li></ul></ul>Ramesh Raskar, http://raskar.info
    5. 5. After X , what is ne X t How to Invent? Ramesh Raskar, MIT Media Lab
    6. 6. X d X++ X X+Y X X ne X t Ramesh Raskar, MIT Media Lab http://www.slideshare.net/cameraculture/raskar-ideahexagonapr2010
    7. 7. Simple Exercise .. <ul><li>Image Compression </li></ul><ul><ul><li>Save Bandwidth and storage </li></ul></ul>What is ne X t Ramesh Raskar, http://raskar.info
    8. 8. Simple Exercise .. <ul><li>Image Compression </li></ul><ul><ul><li>Save Bandwidth and storage </li></ul></ul><ul><li>.. Video Compression .. </li></ul><ul><ul><li>Extend the idea to time dimension </li></ul></ul>Ramesh Raskar, http://raskar.info http://www.slideshare.net/cameraculture/raskar-ideahexagonapr2010
    9. 9. Is project worthwhile? Heilmeier's Questions <ul><li>What </li></ul><ul><ul><li>What are you trying to do? Articulate your objectives using absolutely no jargon. </li></ul></ul><ul><li>Related work </li></ul><ul><ul><li>How is it done today, and what are the limits of current practice? </li></ul></ul><ul><li>Contribution </li></ul><ul><ul><li>What's new in your approach and why do you think it will be successful? </li></ul></ul><ul><li>Motivation </li></ul><ul><ul><li>Who cares? </li></ul></ul><ul><ul><li>If you're successful, what difference will it make? </li></ul></ul><ul><li>Challenges </li></ul><ul><ul><li>What are the risks and the payoffs? </li></ul></ul><ul><ul><li>How much will it cost? </li></ul></ul><ul><ul><li>How long will it take? </li></ul></ul><ul><li>Evaluation </li></ul><ul><ul><li>What are the midterm and final &quot;exams&quot; to check for success? </li></ul></ul><ul><li>Raskar additions </li></ul><ul><ul><li>Why now? (why not before, what’s new that makes possible) </li></ul></ul><ul><ul><li>Why us? (wrong answers: I am smart, I can work harder than others) </li></ul></ul>http://en.wikipedia.org/wiki/George_H._Heilmeier#Heilmeier.27s_Catechism
    10. 10. Great Research: Strive for Five <ul><li>Before Five teams </li></ul><ul><ul><li>Be first, often let others do details </li></ul></ul><ul><li>Beyond Five years </li></ul><ul><ul><li>What no one is thinking about </li></ul></ul><ul><li>Within Five layers of ‘Human’ Impact </li></ul><ul><ul><li>Relevance </li></ul></ul><ul><li>Beyond Five minutes of description </li></ul><ul><ul><li>Deep, iterative, participatory </li></ul></ul><ul><li>Fusing Five+ Expertise </li></ul><ul><ul><li>Multi-disciplinary, proactive </li></ul></ul>Ramesh Raskar, http://raskar.info
    11. 11. Pick atleast 2 out of 3 Fun Cool Media Coverage Impact Money Social implications Research Novelty Generality Science
    12. 12. <ul><li>What distinguishes ML projects .. </li></ul><ul><ul><li>Synthesize not just analyze </li></ul></ul><ul><ul><li>Use power of human intelligence </li></ul></ul><ul><ul><ul><li>Intelligence Amplification </li></ul></ul></ul><ul><ul><ul><li>Human in loop, </li></ul></ul></ul><ul><ul><li>Democratize, Power to the People </li></ul></ul><ul><ul><li>Be paranoid .. Are we relevant and what is next? </li></ul></ul><ul><li>Topics for discussion (create your own group) </li></ul><ul><li>How to pursue 'ideas in the spirit of the media lab'? What is not in the spirit of ML? </li></ul><ul><li>How to make the best of ML resources? </li></ul><ul><li>What are the common problems in picking/initiating/pursuing/finishing great projects? </li></ul><ul><li>Case studies of successful transitions of efforts into research/demos/products and more </li></ul><ul><li>Some procedural topics: juggling classes vs research, Apprenticeship vs independent research, group dynamics, media coverage </li></ul>Ramesh Raskar, http://raskar.info
    13. 13. <ul><li>Abstract </li></ul>Thesis Organization <ul><li>Acknowledgements </li></ul><ul><li>List of Tables and Figures </li></ul><ul><li>Introduction </li></ul><ul><li>Background </li></ul><ul><li>Mechanism/Process Design </li></ul><ul><li>Experimental Methods </li></ul><ul><li>Results </li></ul><ul><li>Discussion and Conclusions </li></ul><ul><li>References </li></ul><ul><li>Appendix A, B, etc. </li></ul>Ramesh Raskar, http://raskar.info
    14. 14. <ul><li>http://www.cs.indiana.edu/mit.research.how.to/section3.11.html </li></ul><ul><li>http://www.cad.auckland.ac.nz/index.php?p=a_good_thesis </li></ul><ul><li>http://www.sce.carleton.ca/faculty/chinneck/thesis.html </li></ul>Ramesh Raskar, http://raskar.info
    15. 15. <ul><li>Be proactive not reactive </li></ul><ul><li>Generalize today’s concepts </li></ul><ul><li>Avoid basing all on facebook/twitter/kinect etc or today’s hot tech </li></ul><ul><li>But we are still slaves to available tech </li></ul><ul><li>“ Let’s do smart things with stupid technology today, rather than wait and do stupid things with smart technology tomorrow” - Bill Buxton. You can ofcourse do even smarter things with smart tech. </li></ul><ul><li>Be prepared but careful on what you do </li></ul><ul><li>Be in optimist but be paranoid (vs pessimist + laidback) </li></ul><ul><li>Defer judgment (don’t dismiss, believe or start instantly on any idea) </li></ul><ul><li>Overnight success after months of work </li></ul><ul><li>Have a list of 10-20 problems .. Don’t work on first one you think or that comes your way </li></ul><ul><li>Talk to a lot of people (you trust) to see if worth purusing as most ideas will be useless anyway </li></ul><ul><li>Don’t be religious, listen to others </li></ul><ul><li>I often find people too much in love with a tiny incremental idea if they came up with themselves </li></ul><ul><li>Try and change what is difficult to some other situation which is easily done but is still important </li></ul><ul><li>Don’t work on same project for 2+ years </li></ul><ul><li>Fail fast </li></ul><ul><li>If u want to win .. Be willing change rules of the game, sometimes the game itself </li></ul><ul><li>Remember the 4Ps and their SEQUENCE </li></ul><ul><ul><li>Projects > Papers > Polished Demo/Prototype > Press </li></ul></ul><ul><ul><li>(Note ‘polished demo’ comes AFTER a paper or some external validation) </li></ul></ul><ul><ul><li>Don’t chase press before you have a serious project that is peer-reviewed or validated (novelty and impact should be already understood) </li></ul></ul>Ramesh Raskar, http://raskar.info