DCT: Dual Channel Training of Action Embeddings for Reinforcement Learning with Large Discrete Action Spaces

TL;DR

This paper introduces a dual channel training framework for action embeddings in reinforcement learning, enabling efficient handling of large discrete action spaces and improving policy learning in noisy environments.

Contribution

The paper proposes a novel encoder-decoder architecture with a dual loss for learning action embeddings that facilitate both action reconstruction and state prediction.

Findings

Outperforms baselines in large action spaces

Produces cleaner, more effective action embeddings

Achieves earlier convergence in policy learning

Abstract

The ability to learn robust policies while generalizing over large discrete action spaces is an open challenge for intelligent systems, especially in noisy environments that face the curse of dimensionality. In this paper, we present a novel framework to efficiently learn action embeddings that simultaneously allow us to reconstruct the original action as well as to predict the expected future state. We describe an encoder-decoder architecture for action embeddings with a dual channel loss that balances between action reconstruction and state prediction accuracy. We use the trained decoder in conjunction with a standard reinforcement learning algorithm that produces actions in the embedding space. Our architecture is able to outperform two competitive baselines in two diverse environments: a 2D maze environment with more than 4000 discrete noisy actions, and a product recommendation task that uses real-world e-commerce transaction data. Empirical results show that the model results in cleaner action embeddings, and the improved representations help learn better policies with earlier convergence.