Warren D. Smith

From Chessprogramming wiki
Revision as of 13:29, 4 July 2020 by GerdIsenberg (talk | contribs)
Jump to: navigation, search

Home * People * Warren D. Smith

Warren Douglas Smith,
an American physicist, mathematician, general theoretical scientist, and in 2005 founder of the Center of Range Voting [1]. He received a double B.Sc. from Massachusetts Institute of Technology in physics and mathematics, and a Ph.D. at Princeton University in the field of applied mathematics on computational geometry motivated by Mesh generation, under Robert E. Tarjan and John H. Conway [2]. He was affiliated with the AT&T Bell Laboratories, the NEC Research Institute, Princeton, New Jersey, where he wrote most of his papers during the 90s, and with DIMACS. His research interest covers the broad range of mathematics, physics, information theory, computer science and politics, amongst other things on algorithms, coding theory, cryptography, combinatorics, computational complexity, game tree search, game theory, computer chess, quantum mechanics, Bayesian statistics, range voting and direct democracy. With the advent of deep learning MCTS chess engines after the AlphaZero publications, Warren D. Smith discussed Leela Chess Zero as well as Stockfish topics in LCZero and FishCooking forums.

Test Suite for Chess Programs

In 1993, along with Kevin J. Lang, Warren D. Smith published and described a test suite for chess programs of about 5500 positions, optically scanned from chess books. The suite is arranged in a number of files, each file being thematic by difficulty, tactical, positional and others [3].

BPIP-DFISA

Along with Eric B. Baum et al., Warren D. Smith worked and published on game tree search, in particular the BPIP-DFISA (Best play for imperfect players - Depth free independent staircase approximation) search control model [4] [5], further generalized in 2004 by Yan Radovilsky and Solomon Eyal Shimony [6].

Smith Notation

Warren D. Smith proposed the Smith notation [7] as used in the Internet Chess Club chess server. The Smith notation encodes moves without any ambiguity using from-square, to-square, and to make it reversible, the captured piece if any - so it is as easy to go backwards as forwards while re-playing a game.

Selected Publications

[8] [9]

1989

1990 ...

1995 ...

2000 ...

2010 ...

Forum Posts

2018

2019

2020

External Links

The Future of Computers, AI (artificial intelligence), etc -- aka, You're gonna die, sucker by Warren D. Smith, March 07, 2013

References

Up one level