Finding a PhD Topic

1
Finding a PhD Topic

• Kathy Yelick
• EECS Department, UC Berkeley
• and Lawrence Berkeley National Laboratory

2
Who am I?

• Why CS?
• Hooked on the first course
• Why a PhD?
• Feeling of where I fit in
• Personal
• BS/MS/PhD from MIT
• Berkeley Assistant Prof. in 1991
• Married Jim in 1993 Megan in 1996 Nathan in
  1998
• Timing Tenure in 1996, Full Professor in 2001,
  Joint LBNL Appointment
• Hobbies
• Skiing, Soccer Mom, (formerly) crew, hiking,
  biking

3
Fear of Topic Selection

• Settling on a PhD topic is often a low point in
  graduate school
• Even for the most successful students
• Why? Because it is very important!
• Youll work on it for a few years in school
• Often will work in the area for years after
• Will define your area for your job search
• But, you can change areas later
• The topic is likely to shift along the way

4
What is a Topic?

• The difference between a project or area and a
  topic
• What is the thesis of the thesis?
• Base on Five Heilmeier Questions
• 1. What is the problem you are tackling?
• 2. What is the current state-of-the-art?
• 3. What is your key make-a-difference concept or
  technology?
• 4. What have you already accomplished?
• 5. How will you measure success?
• Acks
• Based on Pattersons How to Have a Bad Career

5
What to Consider in Choosing a Topic

• What kind of job are you interested in?
• Top 10, teaching, govt lab, industry
• What are your strengths?
• Programming, data analysis, proofs (key insights
  vs. long/detailed verifications)
• What drives you?
• Technology, puzzles, applications
• Practical considerations
• Does your advisor know anything about it?
• Do you (your advisor) have funding for it?

6
Digression Advisors Perspective

• The funding rat-race
• Your write a grant proposal
• You make promises 3 years out
• Not too specific, but specific enough
• It gets funded
• You hire student A to work on the grant
• Student takes an interesting left turn
• Hire student B to finish the grant work
• Write another grant to cover student A

7
5 Ways to Find a PhD Topic
8
1) Flash of Brilliance Model

• You wake up one day with a new insight
• New approach to solve an important open problem

• Warnings
• This rarely happens
• Even if it does, your advisor may not agree that
  its a great idea

9
2) The Apprentice Model

• 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
• Dont work on something you find boring,
  fruitless, badly-motivated,
• Topics can be too close to an advisors interests

10
3) The Phoenix Model

• You work on some projects
• You think very hard about what youve done and
  are doing to look for insight
• Re-implement in a common framework
• Identify an algorithm/proof problem inside
• The topic emerges from your work
• Especially common in systems (the theory
  variation is the stapler model)

• Warnings
• You may be working without a topic for a long
  time

11
4) The Synthesis Model

• Read some papers from other fields
• Look for places to apply insight from another
  field to your own
• E.g., databases to compilers

• Warnings
• You can spend a career reading papers!
• You may not see any useful connections

12
5) The Expanded Term Paper Model

• Take a course in your area or in an area that
  gives you a new perspective
• E.g., theory for systems and vice versa
• Do a project/paper that combines your research
  area with the course
• Low risk topic selection

• Warnings
• This can distract from your research if you cant
  find a related project/paper

13
What to do when youre stuck

• Read papers in your area of interest
• Write an annotated bibliography
• Read a PhD thesis or two
• Read your advisors grant proposal
• Take a project class with a new perspective
• Do some non-thesis work for your group
• Keep working on something
• Get feedback and ideas from others
• Do an internship

14
Dont be Afraid to Take Risks

• Switching areas can be risky
• Move outside your advisors area of expertise
• Dont know the related work
• Starting from scratch
• But it can be very refreshing!
• Recognize when your project isnt working
• Hard to publish negative results

15
Technology And Courage Ivan Sutherland, 1996

• Courage to perceive risk and proceed in spite of
  it
• Research high probability that an attempt will
  fail
• If inadequate courage, Work up courage, reduce
  risk, reduce perception of risk, or dont do it
• External Encouragement (rewards and punishment)
• Deadlines, groups of people, mentors, seminars,
  tenure, taking / teaching classes, starting
  companies, stock
• Self Encouragement
• Getting started warm-up project, break into
  tasks and do 1st one
• To continue refuse to let urgent drive out the
  important
• Rewards
• Thrill of discovery, following curiosity, beauty,
  simplicity, fun
• Acks
• From Pattersons How to Have a Bad Career and
• Sutherlands paper at http//research.sun.com/tech
  rep/Perspectives

16
Its Not About the Topic

• Its about the area
• Is it important? Timely? Jobs in the area?
• And the tools Many researchers have one really
  good hammer
• Use it to solve many problems
• More experienced than others at using it
• Can be a theoretical technique, a software
  system, etc.