Tweet
As part of the AAAI 2012 Robotics Fair, I am playing a demonstration chess match against a fully robotic player running Houdini. The Lego-built robot is perhaps the closest ever to a full implementation of The Turk (sans turban), with visual move recognition and manual execution of moves on a standard-size chess set. It has been designed by a team at Canisius College in Buffalo led by Debra T. Burhans, who was also on the Program Committee of the Educational Advances in AI workshop of the conference.
The match is set for tomorrow (Tue.) 5-7pm at the Sheraton Centre, 123 Queen St. W at University Avenue. We are still unsure about public access to the Robotics Fair, but send me e-mail if you are interested and perhaps I'll know more after returning to Buffalo from NJ tonight. The fair is on page 13 of the conference program (long PDF file); here is a one-page description.
AAAI is considered the world's leading conference on artificial intelligence. Professor Manuela Veloso of CMU, the President of the AAAI organization, will also be taking part in the Robotics Fair.
I have done a Single-PV run of the Canadian Open Premier, and have posted it on my site here. The error shown is unscaled ("raw") error, which I call AD not AE for "average difference". This is not the full Multi-PV run which I use to compute Intrinsic Performance Ratings, and which would take about 20 days for 300 games. Hence it does not really judge skill, but gives some indication of accuracy as well as serving as my scientific control for coincidence with computers---GM John Fedorowicz was the most accurate, and IM Stefan Zierk was the only other to hit 60% matching to Rybka 3.
If any more games are forthcoming, I will run them and update the file. Thanks to all who compiled them---also of the Toronto International and Quebec Open which I will also run.
The match is set for tomorrow (Tue.) 5-7pm at the Sheraton Centre, 123 Queen St. W at University Avenue. We are still unsure about public access to the Robotics Fair, but send me e-mail if you are interested and perhaps I'll know more after returning to Buffalo from NJ tonight. The fair is on page 13 of the conference program (long PDF file); here is a one-page description.
AAAI is considered the world's leading conference on artificial intelligence. Professor Manuela Veloso of CMU, the President of the AAAI organization, will also be taking part in the Robotics Fair.
I have done a Single-PV run of the Canadian Open Premier, and have posted it on my site here. The error shown is unscaled ("raw") error, which I call AD not AE for "average difference". This is not the full Multi-PV run which I use to compute Intrinsic Performance Ratings, and which would take about 20 days for 300 games. Hence it does not really judge skill, but gives some indication of accuracy as well as serving as my scientific control for coincidence with computers---GM John Fedorowicz was the most accurate, and IM Stefan Zierk was the only other to hit 60% matching to Rybka 3.
If any more games are forthcoming, I will run them and update the file. Thanks to all who compiled them---also of the Toronto International and Quebec Open which I will also run.
Comment