2022

History Compression via Language Models in Reinforcement Learning

Paischer, Fabian, Adler, Thomas, Patil, Vihang et al.

Understand

In a partially observable Markov decision process (POMDP), an agent typically uses a representation of the past to approximate the underlying MDP.

  • We propose to utilize a frozen Pretrained Language Transformer (PLT) for history representation and compression to improve sample efficiency.
  • To avoid training of the Transformer, we introduce FrozenHopfield, which automatically associates observations with pretrained token embeddings.
  • To form these associations, a modern Hopfield network stores these token embeddings, which are retrieved by queries that are obtained by a random but fixed projection of observations.

Reading the bibliography…