Feng-hsiung Hsu

Cocreator of the Deep Blue computer

• Born: 1959
• Place of Birth: Keelung, Taiwan (Republic of China)

Primary Company/Organization: IBM

Introduction

Feng-hsiung Hsu is a pioneer in computer science, focusing most of his work on developing a computer that could play chess well enough to defeat a human grandmaster. Hsu and his team from IBM succeeded on May 11, 1997, when their Deep Blue computer defeated reigning world chess champion Garry Kasparov in a series of six matches, by a score of 3½ to 2½. Hsu went on to become the research manager of the Hardware Computing Group for Microsoft's Asia Research Center.

Early Life

Feng-hsiung Hsu was born in Keelung, a major seaport of about 200,000 people, in Taiwan. As a child, he learned to play Chinese chess (xiangqi), Western chess, and Go, and he has stated that Go was the most influential in teaching him strategy. Hsu received his bachelor's degree in electrical engineering from National Taiwan University in 1985; while a sophomore, he became interested in the problem of building a machine that could beat a human grandmaster in chess after reading Peter Frey's 1977 book Chess Skill in Man and Machine. While at National Taiwan University, Hsu became involved with a microprocessor research project, and after hearing a lecture by a representative of Taiwan's Electronics Research and Services Organization, he decided he wanted to study the design of computer chips. After graduation, Hsu spent two years doing his mandatory military service, then traveled to the United States to study computer science at Carnegie Mellon University. At Carnegie Mellon, he found an active community of people interested in computer chess and was particularly influenced by a talk given by Kenneth Thompson of Bell Laboratories about Belle, a chess-playing computer developed at Bell Labs. Hsu continued his interest in developing a chess-playing computer, and he earned his PhD in 1989; his dissertation was titled “Large Scale Parallelization of Alpha-Beta Search: An Algorithmic and Architectural Study with Computer Chess.” While a student at Carnegie Mellon, Hsu won the Friedkin Intermediate Prize for his work on Chiptest, a computer that achieved grandmaster performance in chess, and also won the Mephisto Award in 1990 for his doctoral dissertation.

Life's Work

Hsu began working at IBM in 1989, joining the team that was working on IBM's Deep Thought chess-playing computer. Hsu became the principal designer for the team that developed Deep Blue, the successor to Deep Thought; other members of the Deep Blue team included Chung-Jen Tan, Murray Campbell, Joseph Hoane Jr., Jerry Brody, and Joel Benjamin. Deep Blue (technically, the IBM RS/6000 SP supercomputer) achieved its success not by trying to mimic human thought—which relies on recognizing patterns and solving problems through efficient, heuristic-based searches—but instead by capitalizing on the processing speed and vast storage available in a modern supercomputer. This approach, often characterized as “brute force,” uses the computer's capability to consider all possible moves and the positions they would create and evaluate the strength of each, using the rules programmed into it. The Deep Blue machine that beat Kasparov was able to evaluate 200 million positions per second, double the amount that the previous Deep Blue, which Kasparov defeated, was able to achieve; the improved Deep Blue had a superior central processing unit (CPU).

The first six chess games between Kasparov and Deep Blue had been played in Philadelphia, Pennsylvania, and ended with Kasparov winning 4–2: he won three games, drew two, and lost one. The rematch, which made history as the first occasion in which a computer defeated the best human player in the world, was played in New York City; Deep Blue won 3½–2½, claiming the victory in two games, drawing three, and losing only one. Kasparov accused the Deep Blue team of cheating by using human chess players during the games (by the rules of the challenge, human intervention was allowed only between games, by way of the programmers' further adapting the computer to Kasparov's game) and claimed that the matches were unfair because he had not had access to Deep Blue's recent games. Another, more general criticism of Deep Blue's success was that it required a team of programmers and chess experts to create a computer capable of defeating a single human player, while in a tournament a player would have to defeat multiple opponents with different playing styles.

Although Deep Blue succeeded in defeating the best human chess player in the world, a task once considered impossible, this accomplishment was possible only because of the humans who had programmed it and who had made adjustments between matches, calibrating the computer specifically to defeat Kasparov. Hsu has stated that the true match was not between a man and a machine but between two men (or teams of men): those creating and programming Deep Blue, and Kasparov and his support team. Hsu has also noted that the process of creating a computer to play chess at the highest level demonstrates the usefulness of supercomputers to amplify human intelligence, rather than to replace it—none of the computer scientists working with Deep Blue could play chess at anything like Kasparov's level, but working as a group and with the capabilities of the supercomputer, they were able to defeat him. Hsu also noted that the brute-force approach to computer problem-solving is a good complement to human capabilities, such as pattern recognition and concept formation, and that it makes sense to design computers not to imitate human thought but to perform the tasks (for example, extremely rapid calculation) that a computer can do better than a person.

After Deep Blue's success, Hsu continued to work on creating a better chess chip. When it became clear that there would be no rematch between Deep Blue (or any other computer) and Kasparov, Hsu shifted his interest away from chess-playing computers and left IBM in October 1999. Hsu became the research manager of the Hardware Computing Group for Microsoft's Asia Research Center in Beijing. His group is focused on creating innovative computer hardware that can achieve tasks faster and more efficiently through improvements in fundamental functions of computing through the design of computer architecture and memory hierarchies.

Hsu won the Grace Murray Hopper Award in 1991 from the Association for Computing Machinery (ACM) for his work on Deep Blue. In 1997, Hsu, Hoane, and Campbell received the $100,000 Friedkin Prize from the Association for the Advancement of Artificial Intelligence for creating a chess machine that defeated the human chess world champion. In 2004, he published an account of his signal achievement at IBM, Behind Deep Blue: Building the Computer that Defeated the World Chess Champion.

Personal Life

Hsu received the nickname Crazy Bird while in junior high school, referring both to his eccentric personality and to the fact that feng is a homonym in Mandarin for “crazy.” He continued to use this nickname and the associated initials, C.B., as an adult.

Bibliography

Christian, Brian. The Most Human Human: What Talking with Computers Teaches Us about What It Means to Be Alive. New York: Doubleday, 2011. Print.

Goodrich, Joanna. "How IBM;s Deep Blue Beat World Champion Chess Player Garry Kasparov." IEEE Spectrum, 25 Jan. 2021, spectrum.ieee.org/how-ibms-deep-blue-beat-world-champion-chess-player-garry-kasparov. Accessed 6 Mar. 2024.

Hsu, Feng-hsiung. Behind Deep Blue: Building the Computer that Defeated the World Chess Champion. Princeton, NJ: Princeton UP, 2004. Print.

Hsu, Feng-hsiung. “Cracking Go: Brute-Force Computation Has Eclipsed Humans in Chess, and It Could Soon Do the Same in This Ancient Asian Game.” IEEE Spectrum 44.10 (2007): 51–55. Print.

Hsu, Feng-hsiung. “Oral History of Feng-hsiung Hsu.” Interview by Dag Spicer. 14 Feb. 2005. Computer History Museum. Web. 4 Sept. 2012.

Newborn, Monty. Beyond Deep Blue: Chess in the Stratosphere. London: Springer, 2011. Print.

Rasskin-Gutman, Diego. Chess Metaphors: Artificial Intelligence and the Human Mind. Cambridge: MIT, 2009. Print.