A good analysis of the embeddings we use in the OpenAI Five model: neuro.cs.ut.ee/the-use-of-em…. After analyzing the rest of the model, concludes: “All the strategy and tactics must lie in one place – 1024-unit LSTM.” Pretty remarkable what LSTMs (invented in 1997!) can do at scale.
2
88
4
268
Love this article! I'm curious now, ARE the weights shared between the five agents?
1
2
Replying to @andrey_kurenkov @mitchellgoffpc
Weights are shared. The maxpool happens before any recurrence so it’s just about sharing the inputs, which differ slightly per hero.

Sep 12, 2018 · 4:22 PM UTC
4