Statement of Purpose

Friday, August 7, 2009

Statement of Purpose Computer Science

Statement of Purpose
Computer Science

Like most young kids in computer science, my first interest was in computer graphics. I remember my interest in graphics led me to build a single player Tic-Tac-Toe game with some nice graphics and show it off to my friends. However, I soon realized that my Tic-Tac-Toe is actually quite “stupid” and almost always lost to my friends. I started working on it and made it more intelligent so that it never loses to anyone, and wins most of the times. I didn’t realize that I was working on Artificial Intelligence at that time, but that’s how my interest in AI started growing. Since then my exposures, associations with clubs, readings, and undergraduate work – all have served to push me into deciding to study, research, teach and discover different aspects of AI, particularly those that connect computer science with robotics, and logic-based planning under incomplete knowledge. I am now completely taken in by the challenge of solving complex problems, and making generalizations and relationships that concern writing programs which will make machines think, analyze and learn. AI excites me also because it’s becoming a potpourri of all kinds of knowledge’s. Apart from AI I have also developed a keen interest on Bioinformatics through my undergraduate research work during the last few months. Bioinformatics is an area I didn’t know much about. However, since I started my research on sequence analysis, I have got more and more interested on this area, particularly on processes by which a gene's information can be converted into the functions of a cell.

Life outside classes has been quite exciting for me and I am really enjoying my research work under Dr. Gurpur M. Prabhu. I am currently working on a project on sequence analysis of DNA and then predict functions of the resulting protein. The idea we are working on is to use Conway’s Look and Say Sequence and incorporate some of its properties in the DNA sequence. The first obvious difference between the two is that Conway’s sequence contains primarily three characters (1, 2, 3), whereas there are four characters (A, T, C, G) to deal with in the DNA sequence. I first tried to see whether it’s possible to generate characters other than those three, for example 0 or 4. I proved that it’s not possible for any other character to appear in the successors of the original string, unless it was present in the original string itself. Another interesting property of the Conway’s Sequence is its 92 audioactive elements, just like elements in the normal periodic table and the concept of compound that can be formed using these elements. I am currently working on a program which tests and verifies the similarity of the properties of these elements with the real world elements. Recently I found that though the compound MgCl cannot exist in the real world (only MgCl2 exists), this compound can be formed using Conway’s elements.

I am currently also working on a semester long research project for an AI course on developing artificial intelligence for the game Othello, also known as Reversi. I am particularly focusing on the evaluation function and the search techniques. I also discovered that the mobility of the opponent is crucial to the strategy. The idea is to make a move that will restrict the possible moves of the opponent. I am using alpha beta search for this problem. I discovered that finding the better move towards the beginning cuts off many moves and thus makes it faster. Currently I am working on this algorithm to order the moves so that the better moves come first. I also plan to make it learn inductively in the next step.

Another project I recently started under Dr. Prabhu is on a concept called negative-language. At times it is easier to define something by saying what it’s not than what it is. One such example is justice; it’s easier to say what justice isn’t than what it is. I am using Disjunctive logic programming to work on this problem using DLVk Java Wrapper. Often it’s easier to decide whether a particular solution is optimal or not, than to find the optimal solution. Our goal is to solve this kind of problems effectively.

Now I am also closing in on a yet another watershed decision in my life – of studying higher and deeper into the area of computer science and get into research in fields like AI, or Bioinformatics. Under the Ph.D. program I will get enough scope to translate my thinking and ideas to reality by application to the real problem. This will build up my self-confidence if I am successful. Otherwise, if I fail in a particular application I will get scope to find the reasons and modify it in the next step. Actually, I’m looking forward to an in-depth and challenging hands-on experience in my Ph.D. career. Being on a Ph.D. program will also provide me the opportunity to attend conferences and symposium on my chosen area and I will be able to listen to the experts in the area as well as have interactions with them. This will certainly enable me to enrich my knowledge and keep myself updated on the recent developments. Moreover, I believe that the Ph.D. program will help me to enhance my independent thinking and analysis of a problem. This will be of much importance to me in the next career step, when I aspire to take up my own problem and build up own group.

I believe, doing my Ph.D. in University of California, Santa Barbara will expose me to a deeper area of my chosen topic and will provide me a much optimized environment to expand my knowledge on the subject through associations and interactions with my supervisor and other faculty experts in the department. My reasons of choosing University of California, Santa Barbara are three. First, this university offers a wonderful Ph.D. program in Bioinformatics and Artificial Intelligence; two, it has an acclaimed network of closely working labs that engage in lots of sponsored research Bioinformatics and Artificial Intelligence. On my own, I am interested in doing something in areas like Active Logic, Metacognitive Computation, Simple Hierarchical Ordered Planning, and Sequence Analysis of DNA and all of these are very much there in University of California, Santa Barbara in some variant or other. And third, most importantly, presence of professors like Dr. Ambuj Singh, Dr. Tobias Hollerer, Dr. Terrence Smith, and Dr. Matthew Turk who have been passionately working in the field of Bioinformatics and Artificial Intelligence also influenced me a lot. The presence of these factors led to University of California, Santa Barbara being my preferred choice. If I do my Ph.D. from University of California, Santa Barbara, I will have the opportunity of interacting with them, and who knows, of even working with them on a path breaking idea.

No comments:

Post a Comment