To:  Director of Postgraduate Admissions
Dr. J. J. Bryson

Date:  11 March 2005



Xiao Lu has an excellent CV, and would probably work well at Bath.  However, he has no particular research program that he wants to pursue, except that he loved working with robots on RoboCup.  Also, although his English seems excellent & is very easy to understand, he could not articulate the contributions of his award-winning dissertation in English.  From what little he said I was able to guess that he had derived some of his ideas from Peter Stone & asked him about that.  Xiao Lu seemed very excited that I was familiar with Stone's work, but still could not go much further.

It seems to me that he would be better situated in a US PhD program where he would have a number of years to choose something he wishes to pursue.  Indeed, he says he has applied to several US programs, but was rejected by the best ones because of his GRE marks (presumably, the English one, since his quantitative was exceptionally high -- 790 out of a possible 800.)  However, I still think his profile is promising.  I will write him back and suggest that, if he does not wind up doing a PhD this year, he consider working further on his English and reapply next year.  But given our sparse funding this year, I don't think he should be considered a high priority.

PS I received this email the day after the interview:

Dear Dr. Bryson,

Thank you for interviewing me over the telephone. I am sorry that I did not catch you at the beginning. I have learned much about your work on VR game (UT) and I am quite interested in it.

As you asked about my past research, my main work was focused on the AI part of RoboCup. In our decision-making system, the goal is decomposed to hierarchical sub goals, and sub goal is divided into off-line and on-line parts. Off-line learning is achieved by q learning or fuzzy inference system. On-line planning depends on rule-based system to impose limitations on developed alternatives on-line.

I tried to make a clear idea about what exactly I want to research as a PhD student, but I still do not have the answer. What I am thinking now is to do some deeper research on AI that can be applied to VR character or real robot, and I hope this answer would not make you feel disappointed.

It was great to talk with you Dr. Bryson. Thank you again for your time and please feel free to contact me if you have any further questions.


LU Xiao


Acceptable, but not high priority.  Probably reject with invitation to reapply (unless we suddenly get surplus funding!)