Introduction
What are the Differences between
• Transition from an undergraduate student to a graduate student
Solving a Homework Problem and • How to become an independent researcher?
Solving a Research Problem? • Before graduate school:
• Mostly trained to solve homework problems
Minh N. Do • In graduate school (and beyond, for researchers):
• Ability to solve research problems is the key for success
Electrical and Computer Engineering
University of Illinois at Urbana-Champaign • Analogy: differences between winning a battle and
winning a war!
Solving a Homework Problem
• Example: Show that
is an orthogonal basis for bandlimited functions
“Most advances are made in response to a need, so that it is
necessary to have some sort of practical goal in mind while the • Required skills for solving homework problems:
basic research is being done; otherwise it may be of little value.” • Knowledge
• Creativity
John Bardeen (two Nobel Prizes in Physics) • Persistence
• Typical: Problem ! Solution
• Problem is well defined
• Know that a solution exists!
2 3
Solving a Research Problem
• Example: Can we recover a function from its samples? “There are many things one doesn!t understand and therefore,
we ask them why don!t you just go ahead and take action, try to
• Research problems are typical fuzzy and open ended do something? You realize how little you know and you face
your own failures and you simply can correct those failures and
• Problems are not clear and well-defined
redo it again and at the second trial you realize another mistake
• Problems are not even exist before
or another thing you didn!t like so you can redo it once again. So
by constant improvement, or should I say, the improvement
• Key: being able to ask the right questions and refine them based upon action, one can rise to a higher level of practice and
knowledge.”
• We need not just to solve the problem but also convince the
world that we actually solve it! Fujio Cho (President of Toyota)
• Perform experiments to confirm the theory
• Demonstrate the impacts in real applications and technology
4 5
Five Steps in Solving a Research Problem Define/Formulate a Research Problem
• Step 1: Define the problem (scope, area,…) • Fact: Most of engineering problems are ill-posed!
• Review literature (lot of reading, but not too much)
• Gain deep understanding by doing (implement, experiment,…) • Often we need to redefine/reformulate the problem to make it
• Step 2: Formulate the problem (into a homework problem) solvable
• Find the right model, setup • Imposing realistic assumptions
• Introduce assumptions to simplify the problem • Add constraints
• Step 3: Solve the problem (~ solve a homework problem) • Simplify the problem or model
• Many times need to validate with experiments, applications
• Step 4: Interpret the solution • Engineers: We are allowed to change the problem!
• Go back to the original problem: new insights, new methods,…
• Step 5: Disseminate the results • Many times, being able to define/formulate a research problem
• Papers, presentations, patents,… is half of the work!
6 7
Some Techniques in Formulating Problems Required Skills for Solving Research Prob
• Keep simplifying the problem until it is solvable (e.g. to a special
case) and then try to generalize/extend the solution • Knowledge (much wider)
• Simple problem often gives clear insight and intuition
• Creativity
• Persistence (much more)
• If we cannot solve a simple version of the problem then
we also cannot solve the complex version • If we keep trying hard, something good will come out
• Multiresolution technique: start with a coarse and solvable
problem and successively extend it to make it more realistic • Ask good questions
• Faith (this problem can be solved!)
• Look at the data!
• Flexibility (if not, how can I reformulate it to be solvable!)
• Ask the converse questions
• Communication skills for disseminating results
• Example: we know that bandlimitedness leads to samplable;
• See the the big picture
but what are other samplable signals?
• Organized, motivated, and have a sense of purpose
8 9
Epilog
• Actually, research are typical much more unstructured than what
I just described “Your work is going to fill a large part of your life, and the only
way to be truly satisfied is to do what you believe is great work.
• But a good research should solve a good and relevant problem And the only way to do great work is to love what you do. If you
haven't found it yet, keep looking. Don't settle.”
• Further reading: Steve Jobs (CEO of Apple Computer)
• “How to solve it” by G. Polya, Princeton University Press
10 11
