Search found 8 matches

by Daur
Thu Jan 21, 2021 7:21 am
Forum: MCTS, RL, and neural networks
Topic: TD(1)
Replies: 8
Views: 16250

Re: TD(1)

A challenge for the future would be to try to design algorithms that combine the data-efficiency of local linear approximators with the approximation ability of feedforward networks. You wrote this at the end of your paper ESANN2004.pdf Has progress been made on this topic ? How exactly would you m...
by Daur
Tue Jan 19, 2021 8:46 pm
Forum: MCTS, RL, and neural networks
Topic: TD(1)
Replies: 8
Views: 16250

Re: TD(1)

The successor to RARS is TORCS: http://www.berniw.org/trb/ It is more complex, but also more interesting. I made a TORCS driver during my thesis, too. It was fun. Now that looks exactly like the kind of things that i might try. Thanks a lot ! I'll be sure to keep some update if i'm able to do somet...
by Daur
Tue Jan 19, 2021 6:34 pm
Forum: MCTS, RL, and neural networks
Topic: TD(1)
Replies: 8
Views: 16250

Re: TD(1)

A way to get a recognition for a good machine-learning system is to win competitions. Some don't require that you disclose your code or method. Kaggle has competitions with big money prizes: https://www.kaggle.com/competitions Yes, i did have a look at it. Did you ever win any money out of it, or d...
by Daur
Tue Jan 19, 2021 1:05 pm
Forum: MCTS, RL, and neural networks
Topic: TD(1)
Replies: 8
Views: 16250

Re: TD(1)

Hello there. I don't know if this is the right place to ask my question, but i'll try anyway (feel free to remove this message if it doesn't belong). I have trouble with interacting with people either online or offline, i would often do things the wrong way and make people crazy at me and get banned...
by Daur
Tue Jan 19, 2021 11:38 am
Forum: General Discussion
Topic: Fog of war chess ?
Replies: 2
Views: 4877

Re: Fog of war chess ?

From the little time i spent looking at it, the topic of solving imperfect information games (like poker) is more or less closed. https://www.youtube.com/watch?v=BhUWvQmLzSk Combining Deep Reinforcement Learning and Search for Imperfect-Information Games https://arxiv.org/abs/2007.13544 Starcraft-II...
by Daur
Tue Dec 01, 2020 3:15 pm
Forum: General Discussion
Topic: The Game-AI Forum is back!
Replies: 5
Views: 8890

Re: The Game-AI Forum is back!

Thanks Remi. Actually the fact that the forum is mostly dead, maybe is exactly what i need right now. A place to exist. A fireplace to gather around, and share ideas and old war stories. I don't know you, that well, but you have been a recurring distant internet figure in my life. I believe this for...
by Daur
Tue Dec 01, 2020 10:51 am
Forum: General Discussion
Topic: Fog of war chess ?
Replies: 2
Views: 4877

Fog of war chess ?

Hello there. The chess forum link seems to be dead. I have been playing a few games of "Fog of war" chess lately, and was wondering if anyone has been focused on solving that game. It looks very intriguing from a game theoretical perspective, with both the branching of chess at each game node, and t...
by Daur
Tue Dec 01, 2020 10:42 am
Forum: General Discussion
Topic: The Game-AI Forum is back!
Replies: 5
Views: 8890

Re: The Game-AI Forum is back!

Hello there, i couldn't find any "presentation" section, or really any section for passerby, just wanted to tchat i guess. I'm a bit nostalgic of old days internet, when it was so .. alive. I'm interested in neural networks and general purpose solvers, but not very knowledgeable and with a very poor...