Offline Reasoning for Efficient Recommendation: LLM-Empowered Persona-Profiled Item Indexing

TL;DR

This work introduces Persona4Rec, a recommendation framework that performs offline reasoning to construct interpretable persona representations of items, enabling lightweight and scalable real-time inference and demonstrating that Persona4Rec offers a practical and interpretable solution for next-generation recommender systems.

Abstract

Recent advances in large language models (LLMs) offer new opportunities for recommender systems by capturing the nuanced semantics of user interests and item characteristics through rich semantic understanding and contextual reasoning. In particular, LLMs have been employed as rerankers that reorder candidate items based on inferred user-item relevance. However, these approaches often require expensive online inference-time reasoning, leading to high latency that hampers real-world deployment. In this work, we introduce Persona4Rec, a recommendation framework that performs offline reasoning to construct interpretable persona representations of items, enabling lightweight and scalable real-time inference. In the offline stage, Persona4Rec leverages LLMs to reason over item reviews, inferring diverse user motivations that explain why different types of users may engage with an item; these inferred motivations are materialized as persona representations, providing multiple, human-interpretable views of each item. Unlike conventional approaches that rely on a single item representation, Persona4Rec learns to align user profiles with the most plausible item-side persona through a dedicated encoder, effectively transforming user-item relevance into user-persona relevance. At the online stage, this persona-profiled item index allows fast relevance computation without invoking expensive LLM reasoning. Extensive experiments show that Persona4Rec achieves performance comparable to recent LLM-based rerankers while substantially reducing inference time. Moreover, qualitative analysis confirms that persona representations not only drive efficient scoring but also provide intuitive, review-grounded explanations. These results demonstrate that Persona4Rec offers a practical and interpretable solution for next-generation recommender systems.

Figures (6)

• Figure 1: Comparison between existing LLM-based item rerankers (Upper) and our Persona4Rec (Lower). Persona4Rec shifts LLM reasoning from online inference to offline persona construction, enabling real-time recommendation via lightweight similarity scoring.
• Figure 2: Overview of our Persona4Rec framework. The offline process (Left) constructs item personas and trains a user-persona alignment encoder. The online process (Right) reranks candidates via efficient similarity scoring
• Figure 3: Rating distribution of the two dataset
• Figure 4: Comparison of inference scalability with respect to the number of user samples and candidate items.
• Figure 5: Human evaluation of Recommendation Explainability across compared methods. ($*$: p-value < 0.05)