Grad Studies CS 7001
Popular in Course
Popular in ComputerScienence
This 0 page Class Notes was uploaded by Alayna Veum on Monday November 2, 2015. The Class Notes belongs to CS 7001 at Georgia Institute of Technology - Main Campus taught by Nicholas Feamster in Fall. Since its upload, it has received 13 views. For similar materials see /class/234053/cs-7001-georgia-institute-of-technology-main-campus in ComputerScienence at Georgia Institute of Technology - Main Campus.
Reviews for Grad Studies
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 11/02/15
Finding Good Ideas and Problems Nick Feamster and Alex Gray CS 7001 August 28 2006 Questions What s a good idea What makes a good research problem How to find a good research problem What s a good idea Advance in the field Needs to address some need real problem Need could be immediate or longer term Specific need connected to practice Can address the needs of some specific group May not be immediately obvious today that it s useful Needs to offer new perspectiveangle I should have thought of that strong emotions Market value Can you explain it to mom should be simple and coherent Elegance simplicity generality practicality Opens a new direction Many interpretations applications to other areas not previously thought of What s a Bad IdeaProblem and where to find them Your imagination Comic books cartoons IdeasProblems that are solved Literature reviews are important Those that can t be solved or validated When your advisor tries to talk you out of itwhen should you listen Problem is too hard Can you figure out how to leverage help make some new assumptions new subproblem that people hadn t thought of a new perspective to the problem new technique approach etc Fieldproblem area is dead Don t completely accept this You may know something that your advisor doesn t Find a connection with another field Ask yourself Has the situation changed eg memory cheaper CPU faster etc Problem area is too crowded Can you say something new When to bail out on an idea Solution is impractical What s a good idea Looking backward from n years from now Opens up new thinking Treats fundamental issues better Elegance What Makes a Good Research Problem Problem needs a solutionmust be tractable Solutions need not be complete but they should offer new insight Solution sometimes distills a problem to its essence Problem should be challenging Clearly differentiated from previous work Context is clearly defined The Problem Must be Important How to judge this Matter of taste maybe 7 Requires seeing good and bad to help get a sense Test 1 critical mass Is there some group of people that would benefit from a solution Is it a group you care about Examples network operators citizens in developing countries other researchers etc Test 2 time travel V ll the problem go away on its own eg memory is getting cheaper CPUs faster etc The Problem Must Be Difficult There is no point in solving easy problems Recognize the difference between research and a simple matter of engineering L l r We can t solve problems by using the same kind of thinking we used when we created them The Problem Must Create Knowledge Research the process of creating knowledge Building systems tools etc not research unless you devise new techniques or lessons along the way Good questions to ask Do I know the answer to this before I start Are there places to expect surprises Starting points Hypothesis A new approach Serendipity sometimes takes over Problem Should Have a Solution Solution need not always be complete Opening up a new direction for others to explore Negative results are OK sometimes If you knew you d come up with a positive result then you wouldn t be creating knowledge Prevents others from going down the same path What s a good problem to work on Current hype shortterm Actual impact longterm Question How to measure impact Need some reason you think you can make progress Clear objective and success criterion Secret weapons Bigger brain Mastery of a technique or theory Fresh perspective From another field or culture From much later Where to find research ideasproblems Could draw on ideas from multiple areas New connections between multiple areas that people hadn t thought of before Other people have tried and failed How do you find ideasproblems Frustrations Your own Others how to find out about these Read read read to stay relevant Mailing lists Conferences real world ones in particular Certain news rags Economist Tech Review etc Can often be a good source of ideas for applying one problem domain to another etc How do you find ideasproblems The recent literature The old literature Other literatures Analogies Elegance as a guide Where not to find ideasproblems Your imagination Don t make problems up Plenty of real ones out there The new approach question etc should come from your thoughts The problem itself should not Asking for feedback is key Conference proceedings sometimes The last 80 of papers on a topic solve the last 5 of the problem Often that 5 doesn t matter Engineering or time travel will solve it
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'