Behind AlphaGo

#1

Post by geeharry » Fri, Jan 25, 2019, 8:58 PM UTC

Re: Behind AlphaGo

#2

Post by giacaglia » Fri, Jan 25, 2019, 8:58 PM UTC

It would be interesting to know if there were similar techniques employed for AlphaStar and AlphaGo, with DeepMind's use of LSTMs, etc...

Re: Behind AlphaGo

#3

Post by mercautoe » Fri, Jan 25, 2019, 9:00 PM UTC

giacaglia wrote: post #2

It would be interesting to know if there were similar techniques employed for AlphaStar and AlphaGo, with DeepMind's use of LSTMs, etc...

From my understanding AlphaStar used a LSTM to determine the actions of its agents, but it had a transformer to draw attention to a subset of the inputs

Re: Behind AlphaGo

#4

Post by maximur » Fri, Jan 25, 2019, 9:01 PM UTC

giacaglia wrote: post #2

It would be interesting to know if there were similar techniques employed for AlphaStar and AlphaGo, with DeepMind's use of LSTMs, etc...

It would be nice to get an in-depth overview of the techniques used behind AlphaStar...

Re: Behind AlphaGo

#5

Post by rafaiero1 » Fri, Jan 25, 2019, 9:02 PM UTC

DeepMind reduced the problem to search using the two Deep Neural Networks. MCTS is an old technique but it clearly worked here