LettinGo: Explore User Profile Generation for Recommendation System

TL;DR

LettinGo introduces a flexible, LLM-based user profile generation framework that improves recommendation accuracy by leveraging diverse profiles and direct task feedback, surpassing traditional fixed-format methods.

Contribution

The paper presents LettinGo, a novel framework that generates diverse, adaptive user profiles using LLMs and direct preference optimization, avoiding rigid supervised fine-tuning.

Findings

Enhanced recommendation accuracy with LettinGo.

Profiles are more diverse and contextually aware.

Framework outperforms traditional methods.

Abstract

User profiling is pivotal for recommendation systems, as it transforms raw user interaction data into concise and structured representations that drive personalized recommendations. While traditional embedding-based profiles lack interpretability and adaptability, recent advances with large language models (LLMs) enable text-based profiles that are semantically richer and more transparent. However, existing methods often adhere to fixed formats that limit their ability to capture the full diversity of user behaviors. In this paper, we introduce LettinGo, a novel framework for generating diverse and adaptive user profiles. By leveraging the expressive power of LLMs and incorporating direct feedback from downstream recommendation tasks, our approach avoids the rigid constraints imposed by supervised fine-tuning (SFT). Instead, we employ Direct Preference Optimization (DPO) to align the profile generator with task-specific performance, ensuring that the profiles remain adaptive and effective. LettinGo operates in three stages: (1) exploring diverse user profiles via multiple LLMs, (2) evaluating profile quality based on their impact in recommendation systems, and (3) aligning the profile generation through pairwise preference data derived from task performance. Experimental results demonstrate that our framework significantly enhances recommendation accuracy, flexibility, and contextual awareness. This work enhances profile generation as a key innovation for next-generation recommendation systems.