An Extremely Data-efficient and Generative LLM-based Reinforcement Learning Agent for Recommenders

TL;DR

This paper presents a highly data-efficient RL agent for recommender systems using LLMs, achieving competitive performance with minimal training data and time, by fine-tuning pre-trained models and employing preference-based training methods.

Contribution

It introduces a low-cost, data-efficient RL training approach for recommender systems using generative trajectories and preference optimization techniques with LLMs.

Findings

Generated trajectories match human data in task performance.

DPO agent achieved 19% success rate in under 30 minutes.

Limited training time sufficed for competitive results.

Abstract

Recent advancements in large language models (LLMs) have enabled understanding webpage contexts, product details, and human instructions. Utilizing LLMs as the foundational architecture for either reward models or policies in reinforcement learning has gained popularity -- a notable achievement is the success of InstructGPT. RL algorithms have been instrumental in maximizing long-term customer satisfaction and avoiding short-term, myopic goals in industrial recommender systems, which often rely on deep learning models to predict immediate clicks or purchases. In this project, several RL methods are implemented and evaluated using the WebShop benchmark environment, data, simulator, and pre-trained model checkpoints. The goal is to train an RL agent to maximize the purchase reward given a detailed human instruction describing a desired product. The RL agents are developed by fine-tuning a pre-trained BERT model with various objectives, learning from preferences without a reward model, and employing contemporary training techniques such as Proximal Policy Optimization (PPO) as used in InstructGPT, and Direct Preference Optimization (DPO). This report also evaluates the RL agents trained using generative trajectories. Evaluations were conducted using Thompson sampling in the WebShop simulator environment. The simulated online experiments demonstrate that agents trained on generated trajectories exhibited comparable task performance to those trained using human trajectories. This has demonstrated an example of an extremely low-cost data-efficient way of training reinforcement learning agents. Also, with limited training time (<2hours), without utilizing any images, a DPO agent achieved a 19% success rate after approximately 3000 steps or 30 minutes of training on T4 GPUs, compared to a PPO agent, which reached a 15% success rate.