如何判断一个人是否有潜力

作者: Sanjoy Das, Assistant Professor,　Kansas State University

Ph.D. 1994, Louisiana State University. 
Evolutionary algorithms, neural networks, 
artificial intelligence, discrete optimization. 


International applicants seeking assistantship/admission: please do not 
send me emails blindly without reading this 

Every day I receive emails from international candidates for information, 
for advice, and especially requesting financial assistance. Unfortunately, I 
delete most of them without even reading them carefully. I get so many 
irrelevant emails that cannot reply to each. Extremely rarely has one 
caught my attention. Please do not send me emails without reading this 
first. I am only interested in corresponding to candidates who are 
exceptional, and have a strong interest and experience in my immediate 
area of specialization, and I highly prefer ones with a Master's degree. 

Here are two sample letters: 

A sample letter that will never get the candidate an assistantship with me: 

Unwanted email 

Respected sir, 

I am very interested in going to the US for higher studies. I am seeking 
admission and financial assistance in your esteemed University. I am 
very interested in artificial intelligence and computer architecture. I have 
programming experience in C++, C and ｊａｖａ. I am also familiar with 
several software packages in the Windows environment (MS Word, Excel 
Spreadsheet etc.). I also have networking certification from NOVELL. My 
GRE score is 2050, and I got 99 percentile in the TOEFL exam. 

I request you to spare a few moments from your precious time to go 
through my attached resume. 

Sir - if you are kind enough to offer me assistantship, I promise to work 
very hard for you. I have a very good academic record since primary 
school. Alternately, can you tell me what are my chances of getting 
assistantship from your department? 

Can you kindly inform me about the last dates for Summer term 2002 
before which I can apply? 

Thank you for your valuable time, 

Yours obediently 

ABC 

My comments: I simply delete these emails, which I consider garbage . 
The person seems to be too obsequious and flattering. In the emails that 
I do read (occasionally), I do a first scan for keywords like "precious", 
"respected" etc. deleting the email if I find one. Obviously the person is 
not professional enough. The person is interested in artificial intelligence 
and computer architecture - these are quite unrelated areas, making me 
think that he is trying to cover as many professors as possible to get aid 
from any possible source. Besides, he does not have anything in support 
of his claim (about his interest in artificial intelligence). He emphasizes on 
his software skills, making me think that his goal is to get a programming 
job in the US and higher studies is only to allow him access to the US job 
market. His GRE scores are OK, but not exceptional. He asks me to tell 
him what his chances of getting an assistantship are, elsewhere in the 
department. I have absolutely no way of finding that out, and even if I 
knew, I would not bother to reply back. He also asks me to give him 
information about summer deadlines. How am I supposed to know that? 
(I would typically expect that this information should be available from 
the University web site, or from the grad school, but my guess is as good 
as yours.) Finally, this letter is too generic . He probably uses exactly the 
same format for hundreds of email letters to various professors 
throughout the country, leaving no stone uncovered. I receive hundreds 
of email like this. Regretfully, I cannot answer everyone, and these 
applications do not interest me at all. 



Now, here is a letter that would DEFINITELY attract my attention: 

Excellent email 

Dear Prof. Das, 

I am applying to KSU (amongst other places) for a Ph.D., and I am very 
interested in your research areas. I would like to explore the possibility of 
doing my research with you, with financial support. 

One possible area that would interest me a lot is in applying Tikhonov 
regularization to your problem in computational neuroscience, since I 
understand that neural data can be very noisy, and yet sparse. 
Alternately, I am also interested in applying evolutionary algorithms to 
your genetic neural network parameter estimation. One of my other 
interests in in parallel processing. Are you considering parallel
implementations for any of your optimization/estimation algorithms? 
Considering that the projects that you list in your web site would probably 
be very computationally intensive, that may be a worthwhile strategy to 
pursue. 

I presented a paper entitled "Hybrid Gradient Descent Based Training of 
Probabilistic Recurrent Networks" in the International Conference on 
Neural Networks and Image Processing, Bangalore, last year. It is a 
reviewed publication. If have attached a copy of the article in pdf format. 

If you have available funds, and are interested in my application please 
let me know soon. 

XYZ 

My comments: This is a straightforward letter from someone with a 
genuine interest in my area . He is an exceptional candidate. He has 
evinced a keen desire to learn, and already has considerable mastery 
over my area of research . If I have any funding available, I would 
definitely try my very BEST to get this person . This person may have an 
MS degree, while many applicants have only a BS degree.