Feng-Hsiung Hsu

Feng-Hsiung Hsu lead of the IBM Research team that built Deep Blue. In the October 2007 issue of the IEEE magazine Spectrum Online, Hsu wrote the article Cracking Go of his efforts at the Microsoft Research Center in Beijing China to seed efforts by university students to solve some of the basic problems related to building a championship level Go computer within the decade.

Dr. Hsu stated in the article Nevertheless, I believe that a world-champion-level Go machine can be build within 10 years, based upon the same method of intensive analysis--brute force, basically--that Deep Blue employed for chess. In the article, Dr. Hsu does not discuss these characteristics, among others, that distinguish Go from Chess:

1. the lack of a whole board opening book. Opening books played a crucial role in Deep Blue.
2. the consequences of a local fight upon other fights and the global position. Chess is essentially a single local fight.
3. the lack of a reasonable whole board evaluation function.

Dr. Hsu mentions two challenges the effort faces:

1. the first problem is the tree of analysis.
2. the second problem is the evaluation of the end positions.

To address the first challenge, Dr. Hsu article states that alpha-beta pruning, null-move pruning applied recursively, method of analogy (see article) and newer chip process technology (Deep Blue used 0.6-micrometer CMOS, as of October 2007, 45-nanometer is available) will address the first challenge.

The evaluation of positions may be more difficult. Weaknesses in the evaluation functions of Deep Blue contributed to its defeat in its first match against Garry Kasparov. (See Behind Deep Blue by Feng-Hsiung Hsu, page 186ff.)

Interestingly, as stated in Cracking Go, the expectation that the method of analogy will be significant in developing a Go engine is based upon the idea that each of the many local battles tends to be unaffected by the battles elsewhere. Therefore, the program should be able to graft deep branches--the kind needed to decide life-and-death questions--from one part of the game tree to another.