That's a very good point Veronique. When AI can't give an explanation you can't do a root-cause-analysis when things go wrong - a cornerstone of the LEAN approach in modern business.
Thanks for the insight, Edward (and I was about to plump for "yes"). I'd seen the Starcraft II initiative but thought it was only action-focused.h
I still believe Bridge could be of interest. Like Yu, I'd been fascinated by the communications - how to communicate precisely with limited ...
Thanks for the corrections Qucheng. Like my bidding, my description was muddled. I was assuming it (AlphaGo? DeepMind Labs?) learns from the bbo archives, not screen or youtube videos. But when it plays I was hoping it could play in in a pixel-based way since bbo probably doesn't support ...
Apologies Kit, I have updated the question to explain what DeepMind does. It is not a bridge program. But there is nothing stopping it from teaching itself to play bridge.
What @Jordan describes is one interesting aspect of the challenge. Could it decipher a bidding system from looking at millions ...
@Stefan - I'm hoping that wouldn't be needed.
DeepMind would need to watch the screen (or digest the bbo archives) and learn
a) that there are a limited number of cards in pack (easy)
b) how the scoring works (easy)
c) that it is playing with a partner (trickyish ...
I sort of agree (having "Jack" at home). But the BBO robots are available on the Web, and that would make it easier for the DeepMind to access. My understanding is that DeepMind can now play video games like brickmaster without being taught the rules. So it's more a ...
1. Identify games where chance is an element, but skills prevail.
In some games the order of play is determined by chance, such as by tossing a coin. Baseball, falls into this category and is tolerated by colleges. Skiing perhaps, too, since all slopes are different.
2. Explain how bridge ...