�ݺ�ߣ

All About Research

Adapted from presentations and slides by:

M.J. Harrold - Georgia Institute of Technology
O. Pearce - Texas A&M University
M.L. Soffa - University of Virginia
P. Raghavan - Penn State University
F. Berman - UC San Diego

The CRA-W Grad Cohort for Women Program

What is Research and Why Do It?
What?
 Investigation of a problem in scientific manner
 Discovery of a solution that advances state of
knowledge in areas from theory to algorithms to
prototypes to experimentation to applications

Why?
 Create, have fun, play
 Invent, be on leading edge of discovery, be a
scientist
 Transfer discoveries to benefit society
 Work in interesting and rewarding careers

Beginning with the End in Mind
A PhD graduate should know how to
 select a difficult and interesting research direction (that
matters to someone)
 form hypotheses that can be scientifically evaluated (and
learn how to perform the evaluation)
 conduct research
 report on the results by becoming an effective oral and
written communicator
 obtain funding, run a research group, be an advisor and
mentor to your own students, etc. (so you can continue to
do research, have an effect on others, etc.)
Plan for a research career not a research project!

Approximate Timeline

Getting started
Completing your dissertation
 Choose an area
 Find research Do the research
advisor Write the dissertation
 Begin research Develop long term career goals
Prepare for your career

1 2 3 4 5 6
year
Identify thesis
problem
Choose Ph.D.
committee

Defining your research

Choosing a Research Area

Criteria
 Exciting and interesting area to you
 Important problems in area
 Research type is suitable to you

Ways to identify a research area
 Take courses, attend seminars and colloquia
 Talk to professors, visitors, other students
 Consider both applied and theoretical areas
 Read widely
 Learn about yourself, what you like, etc.
 Solve some research problems

How to choose a Research Advisor
Advisor-advisee relationships are forever!
 Talk to potential advisors and their advisees
 What are their projects?
 How much time do they spend with students?
 Do they have group and/or individual meetings?
 How long to return written materials?
 How much freedom do they give students?
 How long does it take students to finish?
 What is the placement of past students?
 Does the adviser publish a lot with students? What is the
order of names?
 Who presents the papers that are co-authored?
 Do they have research assistantships?
 Do you feel comfortable with this person as your advisor?
 “Try out” an advisor
 Take one of their courses
 Work with them on an independent study

Working on a Research Team (or not)
Why work on a team (in a group)?
 Part of a larger project (less common in more theoretical
areas)
 Research often a collaborative, social process
 Helps you to learn to communicate ideas
 Try new ideas, practice talks, get feedback on
papers, learn to advise other students, etc.

How to work on a research team
 Carve out your problem in the group
 Be generous with giving credit to others, but
 Stand up for your accomplishments
 Your role in the group will change as you progress

Identifying a Thesis Problem
You need to be an expert in your area
Read papers
Use papers’ references to get to original papers
Keep an annotated bibliography of papers, note
Main contribution
Open questions
How it relates to your interests, work

Talk to experts
When they visit UDel
At conferences and workshops
Attend talks, etc.
Carry a notebook to record notes, thoughts, etc.
Question previous works’ assumptions

Identifying a Thesis Problem
You need to consider potential problems

 Consider hot topics carefully
 Thoroughly understand the problem
 Break problem into manageable pieces
 Develop methods that work for you
When to work deeply, broadly; when to put aside
Set aside blocks of time to focus on research
Work consistently on the problem

Persevere—You Will Find a Topic
“ Every morning I would sit down before a blank
sheet of paper. Throughout the day, with
brief interval for lunch, I would stare at the
blank sheet. Often when evening came it was
still empty… It seemed quite likely that the
whole of the rest of my life might be
consumed in looking at that blank sheet of
paper…” (Bertrand Russell, autobiography)

Went on to publish (with Whitehead) the 3-
volume Principia Mathematica
Remember that drive distinguishes the great
scientists (but brains help …)

Really starting your dissertation

A major transition (often the 3rd year)
 Classes are finally done

 Now, you have to define your own research

agenda
 Self-driven schedule

Completing Your Dissertation
The research itself

 Plan your research, and regularly revisit,
reevaluate, and revise these plans
 Work consistently; work very hard
 Take the initiative in your own research
 Usually start out taking advice
 Advisor sets most goals for your work
 Want to end up being a colleague of your advisor

Should know your research area better than anyone
else (even your advisor)

Should be able to defend your work/decisions
 Should set directions for next subgoals

Completing Your Dissertation
showing the world
 Communicate your results along the way
 Identify publishable pieces of your work
 Make appropriate contacts along the way
 Workshops and conferences


Intern or visit with a research group
 Electronic communication with researchers

 Understand “rules” for authorship
 Develop excellent presentation skills (oral
and written)
 Get feedback for your presentations
 Practice, practice, practice
 To your research group

 In your “SIG” group


At conferences

Persevere
You Will Find that Elusive Result
(hopefully …)
“ Every morning I would sit down before a blank
sheet of paper. Throughout the day, with brief
interval for lunch, I would stare at the blank
sheet. Often when evening came it was still
empty… It seemed quite likely that the whole of
the rest of my life might be consumed in looking
at that blank sheet of paper…” (Bertrand Russell,
autobiography)

There are no guarantees -
that is what makes it research!
And, you’ll often have difficult/stressful times

But remember -
drive distinguishes the great scientists

Some Ways to Find a Topic

There is no “one size fits all”

But, here are six common ways to find a topic

1) A Flash of Brilliance
 You wake up one day with a new insight/idea
 New approach to solve an important open problem

 Warnings:
 This rarely happens if at all

 Even if it does, you may not be able to find an

advisor who agrees

2) The Term Project +
 You take a project course that gives you a new
perspective
 E.g., theory for systems and vice versa

 The project/paper combines your research
project with the course project

 Warnings:
 This may be too incremental

3) Re-do & Re-invent
 You work on some projects
 Re-implement or re-do

 Identify an improvement, algorithm, proof

 You have now discovered a topic

 Warnings:
 You may be without “a topic” for a long time


It may not be a topic worthy of a PhD thesis

4) The Apprentice
 Your advisor has a list of topics
 Suggests one (or more!) that you can work on
 Can save you a lot of time/anxiety

 Warnings:
 Don’t work on something you find boring,

badly-motivated,…
 Several students may be working on the

same/related problem

5) 5 papers = Thesis
 You work on a number of small topics that turn
into a series of conference papers
 E.g., you figure out how to apply a technique

(e.g., branch and bound) to optimize
performance tradeoffs

 Warnings:
 May be hard to tie into a thesis

 May not have enough impact

6) Idea From A B
 You read some papers from other subfields/fields
 Apply this insight to your (sub)field to your own
 E.g., graph partitioning to compiler

optimizations

 Warnings:
 You can read a lot of papers and not find a

connection
 Or realize someone has done it already!

�ݺ�ߣ

Research

Recommended

More Related Content

What's hot (19)

Viewers also liked (9)

Similar to Research (20)

Research