π€ AI Summary
To address the misalignment between pretrained multimodal content features and user behavioral preferences in multimedia recommendation, this paper proposes Behavior-driven Feature Adapter (BeFA). First, attribution analysis is employed to uncover information drift and critical feature omissions in content representations for recommendation tasksβa novel finding. Second, BeFA introduces a lightweight, general-purpose behavior-guided feature reconstruction mechanism that dynamically modulates multimodal content embeddings using user behavioral signals, integrating attention-based modulation and cross-modal alignment. As a plug-and-play module, BeFA is compatible with mainstream recommendation models. Extensive experiments across multiple benchmark datasets demonstrate an average improvement of 3.2% in Recall@10 and NDCG@10, significantly enhancing preference modeling accuracy. The code is publicly available.
π Abstract
Multimedia recommender systems focus on utilizing behavioral information and content information to model user preferences. Typically, it employs pre-trained feature encoders to extract content features, then fuses them with behavioral features. However, pre-trained feature encoders often extract features from the entire content simultaneously, including excessive preference-irrelevant details. We speculate that it may result in the extracted features not containing sufficient features to accurately reflect user preferences. To verify our hypothesis, we introduce an attribution analysis method for visually and intuitively analyzing the content features. The results indicate that certain products' content features exhibit the issues of information drift}and information omission,reducing the expressive ability of features. Building upon this finding, we propose an effective and efficient general Behavior-driven Feature Adapter (BeFA) to tackle these issues. This adapter reconstructs the content feature with the guidance of behavioral information, enabling content features accurately reflecting user preferences. Extensive experiments demonstrate the effectiveness of the adapter across all multimedia recommendation methods. Our code is made publicly available on https://github.com/fqldom/BeFA.