A Simple but Effective Elaborative Query Reformulation Approach for Natural Language Recommendation

2025-10-02

AI Summary

Natural language recommendations often suffer from ambiguous or implicit user intents—e.g., broad queries (“cities suitable for youth-friendly activities”) or indirect ones (“cities appropriate for high-school graduation trips”)—posing significant challenges for intent understanding. To address this, we propose a query reformulation method that jointly balances breadth and depth: (1) breadth is enhanced via subtopic expansion to improve semantic coverage; (2) depth is achieved through LLM-powered generative semantic enrichment. This is the first work to explicitly co-model breadth and depth in query reformulation for natural language recommendation. We further introduce three new benchmark datasets specifically designed to evaluate performance on complex, multi-faceted intents. Extensive experiments demonstrate that our approach substantially outperforms existing state-of-the-art methods across multiple natural language recommendation benchmarks, with particularly pronounced gains in scenarios involving intricate semantic structures and implicit user preferences.

Technology Category

• Natural Language Processing: Question Answering
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Applications of semantic technologies for improving search, browsing, recommendation, personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Natural Language (NL) recommender systems aim to retrieve relevant items from free-form user queries and item descriptions. Existing systems often rely on dense retrieval (DR), which struggles to interpret challenging queries that express broad (e.g., "cities for youth friendly activities") or indirect (e.g., "cities for a high school graduation trip") user intents. While query reformulation (QR) has been widely adopted to improve such systems, existing QR methods tend to focus only on expanding the range of query subtopics (breadth) or elaborating on the potential meaning of a query (depth), but not both. In this paper, we propose EQR (Elaborative Subtopic Query Reformulation), a large language model-based QR method that combines both breadth and depth by generating potential query subtopics with information-rich elaborations. We also introduce three new natural language recommendation benchmarks in travel, hotel, and restaurant domains to establish evaluation of NL recommendation with challenging queries. Experiments show EQR substantially outperforms state-of-the-art QR methods in various evaluation metrics, highlighting that a simple yet effective QR approach can significantly improve NL recommender systems for queries with broad and indirect user intents.

Problem

• Addressing dense retrieval limitations with broad and indirect user queries
• Combining query subtopic breadth and depth in reformulation approaches
• Improving natural language recommendation for challenging travel-related intents

Innovation

• Generates query subtopics with rich elaborations
• Combines breadth and depth in query reformulation
• Uses large language models for natural language recommendation