What to Expect for a New-PG to My Group
           ---------------------------------------

Many students ask me what I expect to be a PG in my group.  Usually I
would respond as follows:

* Research Methodology:

  Many students think that having good academic grades would
  automatically make you a good researcher.  This is not always or
  even often true.  Research takes much more beyond good grades; it is
  a matter of creativity, knowledge, critical and independent
  thinking, persistent and complete work, and strong communication
  skills.  Without them, you cannot do good research.

  - Creativity: You are working in a new and emerging field and
    solving some important problem.  You need creativity to formulate
    a problem, bypass its difficulty, narrow the scope of the problem,
    and eventually solve it by creatively integrating the (limited)
    skills/tools you know of.

  - Knowledge: Chances are that most of the easy problems have been
    solved!  You should be curious in knowledge.  Having both breadth
    and depth in your knowledge can help you immensely in solving
    important but challenging problems.  A broad knowledge lets you
    see the field more comprehensively as if using a luminous and
    diffuse light, while equipping with deep knowledge lets you crack
    a problem as if using a laser.

    Research is theoretical and fundamental in nature.  Often times we
    need to derive the performance trends of our algorithms and
    formulate an issue as an optimization problem.  Engaging in CS
    research does not mean that you can throw away your mathematics.
    On the contrary, I expect you to be reasonably good at
    mathematical tools, whether they are graph theory, game theory,
    queuing theory, probability and stochastic theories, LP and NLP,
    optimization techniques, algorithmic analysis, etc.

  - Critical and independent thinking: Research is largely an
    independent pursuit.  You need to be able to evaluate critically
    and independently the strengths and weaknesses of approaches from
    different fronts.  You need to continuously assess both the
    strengths and weaknesses of your and other's approaches so as to
    make sound judgment and continuous progress.  If you are able to
    do that, your approaches are likely to be strong and of impact.

  - Persistent and complete work: Research takes passion and hard
    work.  Producing good and complete results takes much persistence,
    time and efforts beyond simple clever ideas.  Leaving your work in
    the middle without tying its loose ends weakens your research.

  - Strong communication skills: Probably you are the only person in
    the world working on your formulated problem and know its
    solution.  Therefore, you need to communicate well to others what
    you have achieved in English.  This includes situating your
    problem, explaining your approaches, organizing your ideas to put
    things in place and in perspective, and highlighting your results,
    by means of oral presentation and writing.  A good research paper
    should not only have good results, but also get the messages
    across in simple and effective way.


* Working Attitude:

  We are working as a team.  It is better for you to treat me as your
  colleague than an undisputable instructor.  Though I may have more
  experience than you, don't think that I am always right.  You are
  expected to drive your research agenda and take charge of your
  project by moving forward aggressively.  Develop strong leadership
  in your research.  In our meeting, convince me the compelling
  importance of the problem, present to me clearly what you have done
  and explain confidently why you do it the way it is.

  Nowadays research projects are seldom done in an isolated island.
  You are expected to interact regularly with others, especially your
  group members, so that you can help each other.  Show interest and
  engage in their work so as to broaden your horizon, learn from their
  experiences, generate ideas and pursue solutions together. 

  Independent researchers work in flexible hours to get their jobs
  done.  Treat research as your job.  Though there is no hard rule on
  how many hours you need to work each week, I do expect you to take
  research as seriously, if not more, as any other 40-hours-per-week
  jobs.  Therefore, put your hours and efforts into it.  Given that
  different people have different working habit, I do not have a hard
  requirement on your number of hours in office each week.  However, I
  do expect you to come into office/campus at least half of your time,
  the more the better.  This is to allow others to brain-storm with
  you and you to catch up with others.  Be a good team-player.  You
  need to be reasonably responsive to emails during the day-time as I
  sometimes would like to meet with you to brain-storm or to discuss
  issues.

  You may be occasionally asked to help me with some TA work, paper
  reviews, demos, presentations, industrial projects and conference
  organization.  Normally, this will not take you more than 10 hours
  per week.  Most of these tasks are related to your research.
  Therefore, treat them as a learning experience to enrich your
  research program.


* Straight Priority:

  You can graduate only with A in research, not A in courses.  Your
  research should be better than your course grades, i.e., if you have
  good course grades, I expect you to do even better in research.
  Therefore, research should be at your high priority and don't
  compromise it for your course grades.  On the other hand, I
  encourage you to take as many courses as possible relevant to your
  research so as to broaden your knowledge and sharpen your skills.

  Pay attention to the premier conferences and journals, and read
  papers from them.  Learn from them as well.  A good researcher
  publishes in top venues; therefore, submit your papers to them as
  well.  Often times when deadline is coming, you need to work hard by
  putting in more hours.  Take a half-day or full-day break after a
  big deadline as a reward.


* Graduation:

  You need to be passionate towards research in order to do well.  I
  expect good and original work with depth before you can graduate.
  Therefore, do not expect to get your research degree like your
  bachelor degree, which can be attained after a certain fixed number
  of years.  With dedication and hard work, however, you can usually
  graduate in 2 years for MPhil, and 4-5 years for PhD without a
  master degree.

  For those work which you submit to different venues, I appreciate
  your following them up after your graduation.  This means that you
  may need to take some time after your work to revise your reviewed
  papers and do some more simulation runs.  It is often a pity that a
  good piece of work cannot be carried through by the same student to
  its publication.


* Enjoy life:

  Research, despite of its occasional pressure, is overall fun.  It
  creates new horizon of knowledge and advances the state-of-the-art.
  As an old English idiom goes, "All work and no play makes Jack a
  dull boy."  Research is a long process.  I do not expect you to work
  all day all year round.  On the contrary, I encourage you to keep
  pursuing your own interests in your leisure time, no matter they are
  your musical instruments, singing, magic, squash, reading, writing,
  etc.


Gary Chan
Department of Computer Science
The Hong Kong University of Science and Technology
Clear Water Bay
Hong Kong