Transcript Research skills - Australian National University

Research skills
Francesca Rossi
University of Padova, Italy
Research

The process of
searching


carefully
with a method
to answer a question
Skills - Wikipedia

A skill is the learned
capacity or talent to
carry out predetermined results
often with the
minimum outlay of
time, energy, or both.
 Talent or learned
capacity?
Talent or learned capacity?
Most of the skills can be learnt or
improved over time, if one wants
 Some talent is needed, but alone it is not
enough



People with great talent and no skills obtain
much less than what they could do
Not only technical skills
Many types of research

Theoretical


Experimental


Algorithm development, experimental
results
System development


Theorems and proofs
Coding, testing, engineering, …
Basic research skills don’t differ much
Clear about our goal

Our goal is to produce good research results,
to make advancements in science and
technology
 NOT to get a job, or to make more money, …
 If we do good research, these things will
come, but they are not the goal, they are a
side effect
 If we go for the wrong goal, good research will
probably not come, and not even the goal
Skill 1: Curiosity





Intellectual curiosity
How does it work? What
if you change this? Why
did you do this? …
Ask questions to
everybody (and to
yourself) and in all
circumstances
Look at what others do
Can be very helpful in
learning how to
recognize the interesting
questions
Skill 2: Patience



Good research needs time
Don’t expect great results
immediately
 Theorems may be hard to
prove
 Coding may take more
time that initially thought
to be bug-free
 Experiments need to be
tuned to show interesting
results and lessons
Allow ample time to do either
theoretical or empirical
research
Skill 3: Enjoyment





Research has to be
something you like to do
Don’t do it for some other
reason (money, recognition,
fame, jobs, etc.)
If you enjoy it, more chances
that the results will be
interesting, and all the other
things will come
Just relax and have fun
Think carefully if this is what
you like
 If you can imagine a
better life, you should go
and take that life
Skill 4: Problem formulation

A problem should be
formulated in a simple
and crisp way
 Everybody, even nonexpert people, should be
able to understand it
 An unnecessary
complex formulation
may hide the essence of
the problem and also its
solution
 Talk to others, define
your problem to them,
get feedback
Skill 5: Look for important problems





Working only on the details of
your latest propagator or
theorem can be useful for the
next paper
But devote some time also to
think about the larger picture
What are the important
problems in my field?
Read what others work on
Discuss with other students,
your supervisor, anybody you
meet at the conferences
Skill 6: Review existing work





Being able to find and
evaluate previous work
Look for what others
have done before
starting your research
project
It helps defining and
tuning your problem
May give you ideas on
how to solve it
Avoids reinventing the
wheel and wasting time
Skill 7: Problem solving and
adaptability


Find suitable tools to
answer a question or to
solve a well-posed
problem
Be able to adapt
yourself to the context



Turn a problematic
situation into an
interesting research
problem
Read, read, read
Don’t get discouraged, a
solution can always be
found
Skill 8: Self-critical eye

If you are too sure of your
research results, you will never
notice the flaws
 If you are too unsure, you will
never start
 Flaws should be considered and
noted, they may be useful later
to modify the theory to make
space for them


This can be a source of great
results
Have your work read by others
and listen to the comments and
suggestions

Humbleness: there is always
something you can learn from
others
Skill 9: Being able to learn lessons

A system or a tool may be a useful
object, and experiments may be
good, but what is most important is
the general lesson learnt by
developing it




Others (or yourself) can use this
lesson to build different tools or to
define other experiments
Same also for a theorem, but less
crucial since a theorem is already a
way to abstract and generalize what
has been seen in a specific case
Learn also from failures (or from
rejected papers)
Work locally, think globally

Work hard on your specific
propagator, but think about the
general consequences of your
results
Skill 10: Independence and courage





Independent thoughts
 Not just following others’
ideas
Courage to pursue your ideas
 Self-confidence
 Also courage to know
when to stop
Don’t worry to state your
ideas and to be criticised
Better to be criticised than to
be ignored
No need for people who
follow others
Skill 11: Communicating your results




A paper
A talk
A Ph.D. thesis
General advice:


The idea and
motivation is important,
not just the technical
details
The simpler your way
to present your idea,
the more chances
people will pay
attention
Writing a good paper


Papers communicate ideas
The purpose of a paper is to convey your idea

Here is an interesting problem


Here is my idea


Bulleted list of contributions
Once the reader has the intuition, she can follow the details, not viceversa



Giving creadit to others does not diminish the credit for your paper
Simple is better
Be crisp and clear in the introduction


details, evidence for the claims
Here is how it compares to others’ approaches



examples + general case
My idea works


examples + general case
Even if details are skipped or not understood, at least she has something valuable
Start early, not a week before the deadline!
Have the paper read by experts and non-experts

Every criticism is important to improve the paper
Giving a good talk





A taster for your work
Main idea, not technical details
Examples to motivate the problem considered
Answer this question during the talk: If someone
remembers only one thing from my talk, what would I
like it to be?
Open questions, problems in your research


Be open, others may help
Synthesize sentences in bulleted points
 Your work, no matter how brilliant, becomes valuable
to others only in so far as you communicate it to them
Writing a good Ph.D. thesis

Explain clearly your original contribution to
knowledge and science





Question that you want to answer (or problem to
solve)
Why is it important to answer it?
Did you answer it adequately?
Did you make an adequate contribution to
knowledge?
Takes longer than you think


Not just a collection of your papers
More depth than papers
Skill 12: Participate in the academic
system

Write a review



Ethics




Reviewers are there to help
science go forward, just like
authors
… not to help their papers by
killing others!
The most important asset of a
researcher is his reputation
Freedom in what to do, but we
need to have high ethical
standards
Honesty
Have a good attitute!

The advancement of science
is a distributed collaborative
effort
General advice 1: Work hard

Given two people
with equal abilities,
the one who works
10% more will
produce twice as
much
 The more you know,
the more (and faster)
you learn, the more
you can do
General advice 2: Drive and
commitment

You should be
committed to your
research
 This is not a 9 to 5
job
 It should not be a
sacrifice, but an
opportunity to do
something you enjoy
General advice 3: Openess to other
people




Be open to others and to
others’ work
Read a lot
Talk a lot with other
people
Open vs. closed office
doors

People who work with a
closed door produce
more in the short term
but obtain less, and
less interesting results,
in the long term
General advice 4: Collaboration





Most great work comes
out of collaborations with
others
More than one point of
view
Faster definition of the
problem
Solving eased by
resorting to more
available techniques
More fun
General advice 4: Trust your advisor

He is largely judged
by the success of his
students
 He gets great
inspiration by
working with
graduate students
 He can learn a lot
from graduate
students
General advice 5: Be open to many
diverse experiences





Take different courses
Go to summer schools
and conferences
Spend periods in
industry or other labs
One never knows what
can come out
Never a waste of time!
Summary




Many research skills
You can learn/improve
your research skills
You can produce
sustained great research
results if you are
committed, open,
enthusiastic, honest
Work hard and enjoy
what you do
Some useful sources



You and your research, R. Hamming
Technology and courage, I. Sutherland
Basic research skills in computing science, C.
Johnson
 What is research in computing science?, C.
Johnson
 How to give a good research talk, S. L. Peyton
Jones, J. Hughes, J. Launchbury
 How to write a good research paper, S. Peyton
Jones
 How to organize your thesis, J. W. Chinnek
 How to have a bad career in research/academia,
D.A. Patterson
 Ethical constraints, Toby Walsh