How to Succeed in Research: Advice to CS PhD Students
Brad Karp
Completing a PhD is a daunting task: the standard by which thesis research is
judged is high, the burden of innovation in research falls
squarely on the PhD student, and the thesis document itself is usually far
longer and more complex than any written work a PhD student has
produced previously. While most students expect producing a good solution to a
research problem to be hard, many are surprised that
choosing a good research problem and evaluating a solution to it are often
even more difficult.
In this talk, targeted to PhD students in *all* years, across *all* topics in
Computer Science, I will offer HT Kung's wisdom on how to
succeed in CS research, with a few embellishments from my own experience. HT
Kung was a tenured professor of Computer Science at Carnegie Mellon University
through 1992, when he moved to Harvard University, where he remains today. His
advice draws on his experience making seminal contributions across the breadth
of computer science: from theoretical computer science (bounds on complexity
of numerical computations and circuit complexity), to database algorithms
(optimistic concurrency control), to parallel computing hardware (systolic
arrays), to high-speed networks (credit-based flow control).
In this talk, I will address such practical questions as:
- Why is producing a thesis proposal the most crucial (and often most
difficult) part of producing a thesis?
- What are the best known methods for attaining depth in a topic, and
stimulating yourself to generate research ideas?
- What are the most common misconceptions about the nature of research,
and how to succeed at it?
- What are the greatest pitfalls to successfully producing a thesis?
...among many others.
All PhD students are strongly encouraged to attend.
(based mostly on notes provided by HT Kung)
|
