MUSE: A Simple Yet Effective Multimodal Search-Based Framework for Lifelong User Interest Modeling

TL;DR

MUSE is a multimodal search framework for lifelong user interest modeling that effectively combines simple and complex multimodal signals across different stages, improving recommendation performance with minimal latency.

Contribution

The paper introduces MUSE, a novel two-stage multimodal framework that balances simplicity and richness in modeling, and shares large-scale industrial deployment data and practices.

Findings

Lightweight cosine similarity with high-quality embeddings outperforms complex retrieval methods.

MUSE achieves significant improvements in top-line metrics in Taobao advertising.

Open-sourced dataset enables further research on long-sequence multimodal modeling.

Abstract

Lifelong user interest modeling is crucial for industrial recommender systems, yet existing approaches rely predominantly on ID-based features, suffering from poor generalization on long-tail items and limited semantic expressiveness. While recent work explores multimodal representations for behavior retrieval in the General Search Unit (GSU), they often neglect multimodal integration in the fine-grained modeling stage -- the Exact Search Unit (ESU). In this work, we present a systematic analysis of how to effectively leverage multimodal signals across both stages of the two-stage lifelong modeling framework. Our key insight is that simplicity suffices in the GSU: lightweight cosine similarity with high-quality multimodal embeddings outperforms complex retrieval mechanisms. In contrast, the ESU demands richer multimodal sequence modeling and effective ID-multimodal fusion to unlock its full potential. Guided by these principles, we propose MUSE, a simple yet effective multimodal search-based framework. MUSE has been deployed in Taobao display advertising system, enabling 100K-length user behavior sequence modeling and delivering significant gains in top-line metrics with negligible online latency overhead. To foster community research, we share industrial deployment practices and open-source the first large-scale dataset featuring ultra-long behavior sequences paired with high-quality multimodal embeddings. Our code and data is available at https://taobao-mm.github.io.