Unified Generative Search and Recommendation

2025-04-08

AI Summary

Joint modeling of search and recommendation often suffers from a performance trade-off: semantic relevance (for search) and collaborative signals (for recommendation) are difficult to optimize simultaneously. To address this, we propose GenSAR, a unified generative framework that reformulates both tasks as sequence generation problems. GenSAR introduces dual-purpose item identifiers and a task-adaptive prompting mechanism, enabling synergistic fusion of semantic and collaborative signals while achieving task-aware representation disentanglement. Leveraging large language model–driven generative retrieval, multi-granularity item encoding, and hybrid signal fusion for representation learning, GenSAR achieves state-of-the-art performance on both public and industrial benchmarks. It improves search NDCG@10 and recommendation Recall@20 by over 8% simultaneously—significantly alleviating the inter-task performance compromise.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Modern commercial platforms typically offer both search and recommendation functionalities to serve diverse user needs, making joint modeling of these tasks an appealing direction. While prior work has shown that integrating search and recommendation can be mutually beneficial, it also reveals a performance trade-off: enhancements in one task often come at the expense of the other. This challenge arises from their distinct information requirements: search emphasizes semantic relevance between queries and items, whereas recommendation depends more on collaborative signals among users and items. Effectively addressing this trade-off requires tackling two key problems: (1) integrating both semantic and collaborative signals into item representations, and (2) guiding the model to distinguish and adapt to the unique demands of search and recommendation. The emergence of generative retrieval with Large Language Models (LLMs) presents new possibilities. This paradigm encodes items as identifiers and frames both search and recommendation as sequential generation tasks, offering the flexibility to leverage multiple identifiers and task-specific prompts. In light of this, we introduce GenSAR, a unified generative framework for balanced search and recommendation. Our approach designs dual-purpose identifiers and tailored training strategies to incorporate complementary signals and align with task-specific objectives. Experiments on both public and commercial datasets demonstrate that GenSAR effectively reduces the trade-off and achieves state-of-the-art performance on both tasks.

Problem

• Balancing performance trade-offs between search and recommendation tasks
• Integrating semantic and collaborative signals into item representations
• Adapting model to distinct demands of search and recommendation

Innovation

• Unified generative framework for search and recommendation
• Dual-purpose identifiers integrate semantic and collaborative signals
• Tailored training strategies align with task-specific objectives