What it means to be a Ph.D. researcher?
Download
Report
Transcript What it means to be a Ph.D. researcher?
What it means to be a Ph.D.
researcher?
IIT Madras
1st March 2012
Shankar Balachandran
RISE Lab
Dept. of Computer Science and Engineering, IIT Madras
[email protected]
Disclaimer
The views are mine
Not the department’s
Not the institute’s
You may disagree with me
Take things with a tub of salt
These opinions and views were developed over time. I am
unable to attribute any of these to any one person or event
but I am sure several people and external factors shaped
these views.
What is research?
Dictionary Meaning:
diligent and systematic inquiry or investigation into a subject in order
to discover or revise facts, theories, applications, etc
Key things
Diligent
Systematic
Discover or revise
Facts, theories, applications
What does every Ph.D. student aspire
for when they join?
A theorem or law named after them
Hundreds of citations
Every one recognizes you by your name
Even better, just by your face
Will be the smartest one in the planet
Almost a celebrity status
Identify these famous researchers
Andrew Ng
Charles Leiserson
Andrew Tanenbaum
David
Patterson
Raj Reddy
Djikstra
What Can Happen
Your thesis definitely bears your name
Several other authors and co-authors
Stealing your limelight
Google yourself with the hope that you turn up in the first
few links
Keep looking for “who is citing me”
Your advisor may be the only person that recognizes you
What happened in between
Many a slip between the cup and the lip
Picked the wrong school
Picked the wrong area
Picked the wrong problem
Picked the wrong guide
.
.
.
Frustration during both your Ph.D. days and later
Motivation Level of a Grad Student
Why Should One Do Ph.D.?
Can be one of several reasons that you may have
Want to understand things more deeply
Want to make an impact in the society
Want to become a professor
Want to make money
Each one has their own reason
What does your guide (and others,
including yourself) expect from your
Ph.D.?
Original contribution
Leads to the question:
Why should I innovate?
A more deeper and philosophical question
Can’t we just consume and not produce?
Relative Emphasis that We Tend to
Place
Diligent
Systematic
Discover or Revise
Facts
Theories
Applications
Font sizes are proportional to relative emphasis that a typical student places on different things
Ph.D. is a Personal Endeavor
Yes,
You have a guide
You have a lab and lab mates
There is a community which consumes and produces contents
in your field of interest
But
You have a choice over
What you do, how you do it and when you do it
And also what you don’t do.
I see it as a great way to know your strengths as well as
weaknesses
Research Can Be Frustrating
Not all of us are born bright
“I have no special talents. I am only passionately curious.”
Albert Einstein
Being smart is only part of the Ph.D.
Remember the diligent and systematic part?
Superhuman efforts
Be smart
Manage time well
Ask the right questions
Read, write and speak well
My Personal View of Ph.D. research
Ph.D. student is similar to a child
Infant
Unable to move around on your own.
Too many new things – all equally fancy
Toddler
Beginning to walk on your own
Will fall a hundred times
Young child
Can walk steadily but can’t run
Post adolescence
Sprinter, marathoner ….
Where does your guide fit in?
Your research guide is like a parent
Infant
Carry around
Toddler
Hold hands to walk
Young child
If you fall, the parent lifts you up and dusts you
Post adolescence
Dude, you are on your own
Guide/Student Relationship
Both sides can be unreasonable
No parent can expect an infant to run
No adult should expect the parent to carry them
Unfortunately, this mismatch in the expectations becomes a
sore point
Marathoner vs Sprinter
Marathoner
Chug along slowly but steadily
The horizon is distant
Sprinter
Quick burst of energy
Can’t sustain for a long time
Ideally
Ph.D. requires both sprinting and marathon training
Always focus on the horizon
But don’t lose sight of the potholes that you are driving into
My approach to guiding students
Infant
Learn from me
Child
Learn with me
Post-adolescence
Teach me
Research vs Ph.D.
Can’t you do research without a Ph.D.?
Yes
Can you get a Ph.D. without research?
No and aptly so
History of research
Long history, several discoveries and inventions in several fields
History of Ph.D.
Relatively short
How is research awarded?
Imagine what a researcher in the medieval time was looking
for
Money?
Ph.D.?
Most likely to be a personal, academic and scientific endeavor
Somewhere along the line, we have productized research
Papers
Grants
Ph.D. degree
Research Culture in the West
Research thrives on a community
Producers without consumers and consumers without
producers can be a problem
The west has adequate number of both
Not just production/consumption
Peer group to brainstorm and discuss
Get inspirations for problems
Validation
Build systems
Interact with the industry
Learn to ask the right questions and learn to get your answers
reviewed and checked
In India
At most 50 Ph.D.s graduate from CSE departments
Across all IITs + IISc + NITs + Other universities
Serious researchers
A small (but growing) number
In contrast
US has more than 10,000 Ph.Ds in computer science and
engineering
Many are still active in research
Larger community to plug in to
A fresh student starting out in the US already has a better
start
Why Does It Matter?
You are going to publish in conferences where they are trying
to publish
Competitive and trained
Also have English as their mother tongue which can be an
advantage
Makes it quite tilted in their favor
Peter Norvig’s Suggestion
Teach Yourself Programming in Ten Years
Also applies to research
You start out late but sustained effort can get you there
Imagine yourself where you are going to be ten years down
the line
Prepare yourself for that
Key Things in a Ph.D. student’s career
Courses
Qualifiers
Picking the right research problem and areas
Publishing
Giving talks and listening to talks
Thesis Writing
Courses and Qualifiers
Courses
Expand your breadth
Keep you on toes
Take sufficient math courses all throughout
You don’t know when they will come to bite your back
Qualifiers
Necessary evil
Various schools have various mechanisms
I see it primarily as a mechanism to enable early exit
Picking the Right Research Area +
Problems
Early career decision
Has a profound impact
At that time, you can’t see the scope of the area as a Ph.D.
problem
Also, you cannot possibly imagine post ph.d. relevance
In some sense, students trust guides blindly
If you are a new student
Read a lot and ask a lot of questions
No pride to protect at this stage
Don’t hesitate to question established thoughts
Leads to new discoveries
Publishing
Remember, publishing is the way of giving back to the
community
One can question the relevance of
Counts
Venues
Gaming the system etc.
Bottom line
If you have something interesting to say, say it
You should say it – you owe it to the community
Say it well
Sometimes, it is also the means to marking your territory
Talks
Prepares you for something larger than your Ph.D. student
life
You are expected to give talks throughout your career
The more you give talks, the better you get at
Organizing thoughts
Presentation
Keeping the audience interested
Thesis Writing
This is something that you call your own
You should be proud about it later
Requires a lot of effort
Cannot be just the concatenation of papers
Theses do not have page limits
Theses need not assume a lot of familiarity from the reader’s
side
Illustrated examples are okay
This is the stage where your collective wisdom over the years
must show
Shortage of Time
All this is expected to happen in 5 years
What about the 10-year plan?
Research does not stop with your Ph.D.
Your Ph.D. research should lay the foundation for your research
career
Ph.D. is the path not the goal post
The scenery is as beautiful as the destination
Post Ph.D.
Academic vs Industrial Research
Industrial Research vs. Academic
Research
Academic research can do with hot-potato approach
If it is too hot to handle, drop it
Industrial research always has limited funding
Industrial research can always get axed suddenly
Industrial research is typically more focused and the
timelines are usually rigid
Some Useful Tips for Ph.D. students
Maintain a research scrap book
Could be a notebook or a blog
Note down all your thoughts, questions, methodologies etc.
Can come in quite handy to look at the evolution of your
thought process
Also makes it easy when you write papers/thesis
Keep on the look out for interesting problems
You are going to be employed 5 years from the start of your
Ph.D.
You should think about what is going to be relevant five years
down the line and acquire sufficient skills
Tips (contd.)
In your third year or so
Start thinking about where you want to be
Academia
Start thinking about sending submissions to journals/top conferences
What kind of courses will you be asked to teach
What area of research you want to work after Ph.D.
How to write proposals / get grants
National vs International prominence
Industry
What kind of position are you going to like?
Research vs Development
Taking up Post Doctoral Positions
Very common in sciences, humanities and in engineering
fields other than CSE and EE
Usually done to
Stand on your own feet
Move to a new area
Pick up other skills that you may need to sustain a research
career
Importance of Networking
We all fit in to a community
Start networking with your community right from your Ph.D. days
With people from your research lab and the department
With researchers outside the institute
You could even think of doing independent research or work with
people other than your guides
When you finish your program
You will have your own network
Your own set of research problems
You don’t have to be under the shadow of your guide
Attend conferences, give talks, attend talks by people from other places
Some Useful Resources
How to be a Good Graduate Student by Marie desJardins
A graduate school survival guide: So long, and thanks for the Ph.D.
by Ronald Azuma
Douglas Comer’s essays on Ph.D. in Computer Science
http://www.cs.purdue.edu/homes/dec/
Measuring research
http://www.cs.purdue.edu/homes/dec/essay.research.measure.html
You and your research by Richard Hamming
Thank You