Enhancing Prediction Models with Reinforcement Learning

TL;DR

This paper describes Aureus, a large-scale news recommendation system that leverages reinforcement learning, multi-armed bandits, and large language models to improve personalization and adapt to dynamic news content, achieving significant online performance gains.

Contribution

It introduces Aureus, a novel news recommendation system integrating reinforcement learning with advanced models, demonstrating real-world improvements in online metrics and adaptability.

Findings

Significant online performance improvements with Aureus

Effective handling of cold start and content freshness issues

Successful deployment in a real-world environment

Abstract

We present a large-scale news recommendation system implemented at Ringier Axel Springer Polska, focusing on enhancing prediction models with reinforcement learning techniques. The system, named Aureus, integrates a variety of algorithms, including multi-armed bandit methods and deep learning models based on large language models (LLMs). We detail the architecture and implementation of Aureus, emphasizing the significant improvements in online metrics achieved by combining ranking prediction models with reinforcement learning. The paper further explores the impact of different models mixing on key business performance indicators. Our approach effectively balances the need for personalized recommendations with the ability to adapt to rapidly changing news content, addressing common challenges such as the cold start problem and content freshness. The results of online evaluation demonstrate the effectiveness of the proposed system in a real-world production environment.