Blockwise Sequential Model Learning for Partially Observable Reinforcement Learning
This addresses the problem of inefficient sequential information compression in partially observable settings for reinforcement learning practitioners, though it appears incremental as it builds on existing self-attention and importance sampling techniques.
The paper tackles partially observable reinforcement learning by introducing a blockwise sequential model that processes multiple timesteps per block using self-attention, and it significantly outperforms previous methods in various environments.
This paper proposes a new sequential model learning architecture to solve partially observable Markov decision problems. Rather than compressing sequential information at every timestep as in conventional recurrent neural network-based methods, the proposed architecture generates a latent variable in each data block with a length of multiple timesteps and passes the most relevant information to the next block for policy optimization. The proposed blockwise sequential model is implemented based on self-attention, making the model capable of detailed sequential learning in partial observable settings. The proposed model builds an additional learning network to efficiently implement gradient estimation by using self-normalized importance sampling, which does not require the complex blockwise input data reconstruction in the model learning. Numerical results show that the proposed method significantly outperforms previous methods in various partially observable environments.