Recommendations by Concise User Profiles from Review Text

TL;DR

This paper introduces CUP, a lightweight framework that creates concise user profiles from review texts to improve recommendations for users with sparse interactions, reducing noise and computational costs.

Contribution

It proposes a novel method for extracting informative user profiles from reviews to enhance recommender systems, especially for users with limited interaction data.

Findings

Fine-tuning small language models with constructed profiles outperforms LLM-generated rankings.

The framework effectively reduces noise and computational costs in user profile creation.

Experiments demonstrate improved recommendation accuracy for sparse-user scenarios.

Abstract

Recommender systems perform well for popular items and users with ample interactions (likes, ratings etc.). This work addresses the difficult and underexplored case of users who have very sparse interactions but post informative review texts. This setting naturally calls for encoding user-specific text with large language models (LLM). However, feeding the full text of all reviews through an LLM has a weak signal-to-noise ratio and incurs high costs of processed tokens. This paper addresses these two issues. It presents a light-weight framework, called CUP, which first computes concise user profiles and feeds only these into the training of transformer-based recommenders. For user profiles, we devise various techniques to select the most informative cues from noisy reviews. Experiments, with book reviews data, show that fine-tuning a small language model with judiciously constructed profiles achieves the best performance, even in comparison to LLM-generated rankings.