Representation Learning in Low-rank Slate-based Recommender Systems

TL;DR

This paper introduces a sample-efficient representation learning algorithm for slate-based recommender systems modeled as low-rank MDPs, aiming to improve reinforcement learning efficiency in large state-action environments.

Contribution

It proposes a novel RL algorithm tailored for low-rank slate recommendation environments and constructs a simulation setup for evaluation.

Findings

Effective in large state-action spaces

Improves sample efficiency in RL for recommendations

Provides a new simulation environment for testing

Abstract

Reinforcement learning (RL) in recommendation systems offers the potential to optimize recommendations for long-term user engagement. However, the environment often involves large state and action spaces, which makes it hard to efficiently learn and explore. In this work, we propose a sample-efficient representation learning algorithm, using the standard slate recommendation setup, to treat this as an online RL problem with low-rank Markov decision processes (MDPs). We also construct the recommender simulation environment with the proposed setup and sampling method.