Author: Wayne Lowrance
Date: 16:39:16 10/15/99
Go up one level in this thread
On October 15, 1999 at 16:21:22, Scott Gasch wrote: >Hi, > >Today Dr. Hsu (who recently left IBM / Deep Blue team) gave a talk at Microsoft >research. I went to hear what he had to say and ask him a few questions. > >The presentation was pretty much about the history of computer chess. Bob he >mentioned you. Also Belle and Ken Thompson. And then went on about his own >work on Deep Thought / Deep Blue. > >The entire time I had the feeling he was selling something. Towards the end he >made a pitch for (private) support of his new effort which, as best I can tell, >is a single processor chess card for a normal PC to take on Kasparov for the >world title. He seems to think that Kasparov would put the title on the line in >a rematch. Hsu just left IBM last week and has done no work on this yet but he >is looking at a timeframe of 2001-2002 for the match. Dr. Hsu believes he would >have a better than 50% chance of winning this match. He has also gotten >permission from IBM to start with the deep blue chess chip and build upon it. > >I was able to ask some questions during and after the talk that may be >interesting to you as well. Here is what I learned: > >1) His new project will optionally use NULL move pruning and other pruning. He >wants to be able to turn it on and off. He believes that with proper selective >extension you do not need NULL move pruning at all if you have enough speed, >though. > >2) The eval hardware on Deep Blue was 2/3 of the chip and his eval function has >about 3000 terms which they tuned by hand. He is thinking about different ways >to automatically "train" the eval function now. > >Scott Geeze, at last, terrific for the first time we have sort of heard from Dr Hue. Thank you for attending and posting Wayne
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.