Author: Miguel A. Ballicora
Date: 15:57:05 05/17/02
Go up one level in this thread
On May 17, 2002 at 17:13:38, Russell Reagan wrote: >On May 17, 2002 at 15:39:34, Miguel A. Ballicora wrote: > >>Innovative approaches and breakthroughs are NOT found in the private or >>commercial areas (with few exceptions). They are found in Universities or funded >>by governments or non-for-profit foundations. Private companies do not have the >>money or the time to think "100% free". This is coarse generalization but it is >>quite close to the truth. >>I would not be surprised that "planning in games" comes from an University >>rather than from a commercial programmer. It might not be in a form of >>alpha-beta though, who knows?. You can expect that later this idea is improved >>dramatically by the commercial guys in little time. That is the way basic and >>applied sciences work. >> >>Regards, >>Miguel > >My point was that this is a question that has been basically unanswered because >we simply don't know the answers, that is, "we" the general public do not know >the answer. He seems to have the attitude that he's going to ask this question >and someone is going to come out and say, "Oh yeah sorry about that guys, I've >known about this for years but I just never told anyone. I was just waiting for >someone to ask, so here ya go, here's everything you need to know! Enjoy!" and >that's not going to happen. I mean either no one knows the answer to his >question, or someone does and they haven't shared it with anyone yet, and they >certainly aren't going to start because someone asked on CCC. > >My comments about the answer comming from the "commercial" sector was slightly >misinterpreted by you. When we speak of "commercial" programs, we're talking >about Fritz, Tiger, Shredder, Junior, Rebel, etc. etc., and the truth of the >matter is that in the "commercial" computer chess world, being a "commercial" >program doesn't mean the same thing as it does in other areas of the world. Your In a small scale, I think it does. >point is usually correct. Microsoft probably isn't going to make any great >scientific breakthroughs, and the ones they do aren't very significant when >compared with the academic world's breakthroughs. The commercial computer chess >scene is different however. Most of the authors of the commercial chess programs >started off as amateur programmers. They made breakthroughs (which they >obviously aren't telling anyone and everyone about, which is perfectly >understandable) and since they made breakthroughs, they eventually became >commercial programs. So those guys have a track record for making these kinds of >breakthroughs, and maybe even more so than someone at a university AI department I am not sure about that. I do not think that I agree with you about what a breakthrough is. To me, breakthroughs are, alpha-beta search, PVS search, MTD(f)hashtables (refutation and transposition), zobrist keys, killer tables, history heuristics, null-move, endgame tablebases, nullmove, parallel algorithms (several) etc. Somebody else can confirm this but IIRC most of the original ideas were not born in the commercial area. >would. I don't see any program from MIT beating the daylights out of Fritz or >Tiger. Yes, Belle, Cray-Blitz, Deep Thought, HITECH. Later, those projects lost interest for many reasons (except Deep Blue that turned "commercial" but was born in a University), that is why you do not see that now very often. Anyway it does not matter. In general, you are not going to see a product of the same quality developed in an University. People on research do not care about the perfect product, they care about _proving_ that the idea will work and that the idea is new. Private sector or applied scientists optimize it to perfection. The main goal of a University project sometimes is not winning it all! And do not forget, before the Fritzes and the Tigers there were 20-30 years of academic research in the area. >The other side of the coin is Chinook, developed at a university, and able to >beat all but one human. Okay, maybe technically Chinook beat Tinsley, but it >wasn't in his prime and it wasn't exactly a well earned win since it came from a >forfiet. Nevertheless, he seems to have been the only opponent (human or >computer) that stood a chance against the machine. > >So I'm not saying that university studies won't produce something similar in >chess. But in computer chess, these kinds of breakthroughs thar are not typical >in the regular commercial world are certainly reasonable possibilities. >In any case, you're going to either find what you're looking for in some AI >journal after a lot of digging for what you want, or you're going to have to >figure it out on your own. That's what I think anyway. Yes. My point is that if we are going to find it in the future, I bet it will be in a journal. Regards, Miguel >Russell
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.