Neural nets and dataflow matrix machines
Mar. 30th, 2016 09:40 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
dmmWhat we see is an explosive growth in the capabilities of various deep and/or recurrent neural nets. The latest AlphaGo triumph underscores that...
A couple of particularly notable URLs:
"The Unreasonable Effectiveness of Recurrent Neural Networks" by Andrej Karpathy:
http://karpathy.github.io/2015/05/21/rnn-effectiveness/
"Giraffe: Using Deep Reinforcement Learning to Play Chess" by Matthew Lai:
http://arxiv.org/abs/1509.01549
It is interesting that dataflow matrix machines are a rather powerful generalization of recurrent neural nets.
A couple of particularly notable URLs:
"The Unreasonable Effectiveness of Recurrent Neural Networks" by Andrej Karpathy:
http://karpathy.github.io/2015/05/21/rnn-effectiveness/
"Giraffe: Using Deep Reinforcement Learning to Play Chess" by Matthew Lai:
http://arxiv.org/abs/1509.01549
It is interesting that dataflow matrix machines are a rather powerful generalization of recurrent neural nets.
