Moses Charikar comments

I was fortunate to have written an NSF proposal previously with my colleagues, Sanjeev Arora and Amit Sahai – watching your more experienced colleagues in action is a great way to learn. I was also very grateful to have had several examples of CAREER proposals to look at while writing mine. In particular, I was quite influenced by Salil’s very well written proposal and borrowed several organizational ideas from it.

Some observations:

  1. It is interesting to go back in time to 2002 when this proposal was written. The community has made progress on many of these problems since.
  2. My reviews said that the proposal looked like a laundry list of problems in approximation algorithms and I agree. I pretty much listed every thing I had ever considered working on. In retrospect, perhaps I should have been strategic in picking a coherent subset and leaving material for other NSF proposals down the line. (I was less motivated to do this because I had another NSF grant already). On the other hand, writing the proposal was a very valuable exercise because it helped organize my thoughts on what problems I wanted to work on.
  3. Where possible, I tried to describe half baked ideas and partial approaches for the problems I listed e.g. asymmetric k-center (where I was proved wrong very soon after). This builds credibility by showing that you have spent some time actually thinking about the problems and didn’t just pick them off of an open problem list.
  4. In writing the education plan, it was helpful to have spent a year at Princeton before this was written. This meant that I had actual courses I had taught and knew I would be teaching in the years to come.
  5. I have learnt many useful proposal organization devices in the years since (from collaborations with my colleagues – Avi, Sanjeev and others at the Intractability Center). In hindsight, I wish I had used them in writing my CAREER proposal. For example, it is useful to highlight ideas and techniques that cut across the various problems you discuss – this helps make the proposal more coherent. Also it is is useful to highlight specific research challenges, e.g. (in bold) “Challenge: Exploit Lovasz-Schrijver lifting to obtain better guarantees for graph coloring”. This is a good way to catch the reviewers’ attention as they skim the proposal in the 1 hour or so they have allotted to reading it.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: