Language models improve personalized explanations in recommendation systems
Learning to Rank Explanations for Recommendation with Language Models
Information Retrieval
Summary
It's important for online recommendation systems to give users clear and relevant explanations for their suggestions. The authors propose a new method that uses language models to rank short review-based explanations based on how well they match a user's past interests and the item. Their system learns to find the most personalized and factual explanations, performing better than previous methods in tests on Amazon and Yelp data. This approach helps build more understandable and trustworthy recommendation systems.
What this means in practice
- •For e-commerce engineers: Enhance product recommendation explanations by selecting personalized, factually grounded review snippets using adapted language models.
- •For customer experience teams: Improve transparency in online reviews and suggestions by deploying personalized explanation rankings that boost user trust and engagement.
Authors
Ben Kabongo, Vincent Guigue
Abstract
Recommender systems are ubiquitous on the Web, making understandable explanations essential for users. Although large language models (LLMs) increasingly generate natural-language justifications, evaluating their factual grounding and personalized relevance remains challenging. Statement-level ranking offers an alternative by selecting concise, review-derived explanations, enabling fine-grained and reproducible evaluation using established ranking metrics. However, existing statement ranking methods largely rely on collaborative representations, while semantic enhancements often fail to outperform simple user-history popularity baselines. Moreover, directly adapting language models for personalized statement ranking remains largely unexplored. In this work, we introduce LMStaR (Language Models for Statement Ranking), a fully text-based framework that formulates personalized explanation ranking as dense retrieval. A query encoder represents the user's interaction history and target-item metadata, while a statement encoder maps candidate explanations into a shared embedding space. Both encoders are jointly adapted through multi-positive contrastive learning to learn personalized relevance. Experiments on Amazon Beauty, Amazon Toys, and Yelp show that LMStaR achieves the highest NDCG@1 across all three domains, outperforming CLER by 1.06--8.25 points (+11.8--65.9%). It also leads top-five ranking on both Amazon datasets, although CLER remains stronger on Yelp. Task-specific adaptation improves all five evaluated backbones and transfers across domains without target-domain fine-tuning. Counterfactual analyses further reveal the importance of user-specific histories. These findings highlight the potential of task-adapted language models for personalized explanation ranking.