Masaru Tomita: From game programming to discovering the mystery of life

By admin,

  Filed under: Artificial Intelligence, Biomimicry, Computing, Education, Electronics, Math, Neuroscience, Research, Science
  Comments: Comments Off on Masaru Tomita: From game programming to discovering the mystery of life

ORIGINAL: Keio University
By TOMITA, Masaru Professor

Faculty of Environment and Information Studies

Those college days when I made money by game programming

Masaru Tomita

When I was an undergraduate student, all second year students were expected to take the programming course, which I did not find interesting. For example in class, we had to use Newton’s method to derive a circle ratio “π”, but I was not so excited about using programming method to obtain an answer we already knew. At that time the Space Invaders game was extremely popular. The game was designed in such a manner that if the player cleared all the stages, the game would start all over again. I became so skillful at clearing the stages that the game never ended. Then I thought it would be fun to change the rules to make it more difficult. Thus I started to self-study programming and create my own games. I made Breakout or similar games and sold them to shops in Akihabara Electric Town. At times these games took three days and nights to design, and I could sell each game for about 800,000 yen. It was quite good earning for a university student like me.

I sensed the limits of computer programming

Add caption

No longer satisfied with creating action games and making money, I became interested in intellectual games. Being a dan holder in shogi or Japanese chess since junior high school, I decided to develop a computer game of shogi. Programming the board and displaying the movement of pieces that followed the rules were quite easy. However, I soon realized that programming a string of thought process into the computer to allow it to play against human beings was quite daunting. In the game of shogi, a player has approximately 30 possible moves per phase. This means that a player has to come up with a cube of 30 moves in order to speculate three phases ahead. Even some shogi beginners can speculate 15 phases ahead in a crucial stage, and this means they can plan 30 to the power of 15 moves ahead. Even supercomputers would take an extraordinary length of time to scan all these patterns and choose the best option.

Encountering an academic field called artificial intelligence

Add caption

In reality, human players do not go over all options before choosing the best to make a move. Human mind quickly discards options that are useless, selects a few good moves and chooses the best. In other words, humans proficiently prune the search tree in order to make a selection. I asked my seminar professor how to program a computer to prune the search tree. He told me that the procedure is taught in a field called artificial intelligence. He revealed to me that no Japanese laboratory has yet specialized in artificial intelligence, and if I was serious about studying the field, then I should go to the United States. I did some research about my options and decided to study under Professor Herbert Alexander Simon, the winner of the Nobel Memorial Prize in Economics for research in artificial intelligence (more accurately, decision-making), at the School of Computer Science in Carnegie Mellon University.

Understanding the limits of computers, I sensed the mystery of human life

Though I faced many hardships, I finally managed to acquire a master’s degree and a doctoral degree from Carnegie Mellon University. Afterwards, as Professor Simon’s Research Associate and then as an Associate Professor, I continued to conduct research mainly in machine translation. With the establishment of SFC, I returned to Japan but continued my research in machine translation. The task of designing a machine translation system is quite difficult. Incorporating grammatical rules and dictionary functions into a computer may give it the ability to perform a literal translation, but still it will not be able to match the skill of professional translators. Understanding the meaning of sentences, and amalgamating common sense with human knowledge are necessary to make proper translation. In short, a programmer has to program all aspects of human knowledge into a computer in order to translate properly. This procedure is as difficult as creating the human brain. It is rather difficult for a computer to do things that the human brain can do easily. On the contrary, a human brain, starting with a fertilized egg, develops automatically in about five years through repeated cell divisions. This is quite an amazing process!

Becoming a medical student while being a teacher

The more I studied artificial intelligence, the more I became interested in human mechanism and its ability to create an intelligent system through cell division. This was the time when researchers around the world had just started analyzing the human genome, which is comprised of the letter string ATGC. Since the object of machine translation was also a series of letter strings, I thought I could apply my research findings to genome analysis. The only hitch was that in order to research in this field, I needed knowledge of molecular biology. I felt that if I had to acquire this knowledge, I should become a student of the Graduate School of Medicine. My only concern was that if I started studying biology while being a computer science teacher at SFC, I might not be able to show any substantial research results in a short time. I took my concern to Professor Hideo Aiso, then Dean of the Faculty of Environment and Information Studies. He advised me: “Don’t get obsessed with getting immediate results and confining yourself to one area alone. It is vital for a professor at SFC to open the way in new research fields.” The late thirties of a researcher is an important stage in his life. With Professor Aiso’s encouraging words during those crucial years, I overcame the fear of leaving a blank period in my research career. I then decided to wear two hats for four years, one as a professor at SFC and the other as a student at the Graduate School of Medicine, Keio University.

Study together, search together

Since my enrollment as a student at the Graduate School of Medicine, biology has become the theme of my seminar at SFC. Whereas professors would generally teach a field after having explored it themselves, my policy is to study and conduct research together with my students. SFC offers a unique environment to allow this to happen. Analyzing the genome string with a computer revealed to me many secrets of life. It is exciting for me to discover something unknown, things that are not in textbook, and things that nobody in the world knows about. It is also exciting for students to experience these things together. SFC offers seminars in various fields so that students do not need to limit their interest to a single area of study upon enrollment. They can participate in seminars from their first year and shift into other areas later on, if their subject of interest changes. This allows students to expand their investigative curiosity and discover their own research themes. SFC is an ideal university campus for students with such an ambition.

A Brief Background of Professor 

TOMITA, Masaru
Professor Tomita graduated from Keio University with a bachelor’s degree in Mathematics in 1981. After his graduation, he enrolled in Carnegie Mellon University and acquired a master’s degree in 1983 and a Ph.D. in 1985, both the degrees were in Computer Science. He conducted research at the School of Computer Science, Carnegie Mellon University, as a research assistant. He was promoted to the position of a Research Associate (1985) and then an Associate Professor (1987). From 1987 he has held the position of Associate Director at the Center for Machine Translation. In 1990, Professor Tomita returned to Japan and took a position of an Associate Professor at the Faculty of Environment and Information Studies, Keio University, initially established when the Shonan Fujisawa Campus was opened. He was appointed Professor in 1997. While he was teaching, he acquired a Ph.D. in Electrical Engineering from the Graduate School of Engineering, Kyoto University (1994), and a Ph.D. in Molecular Biology from the Graduate School of Medicine, Keio University (1998). He was the Dean for the Faculty of Environment and Information Studies from 2005 to 2007, and has been the Director of the Keio University Institute for Advanced Biosciences since 2001.

Major Honors:

  • Presidential Young Investigator Award from the National Science Foundation, USA(1988);
  • IBM Japan Science Prize(2002);
  • Minister of Science and Technology Policy Award in recognition for industry-academia-government collaboration performance(2004);
  • Prize for Science and Technology of the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology(2007)

Faculty Profile

Laboratory for Advanced Biosciences, Keio University

Institute for Advanced Biosciences, Keio University

(22 May 2009)Archive

Comments are closed for this post.