DIAURec: Dual-Intent Space Representation Optimization for Recommendation

2026-04-10

AI Summary

This work addresses the challenge that sparse interactions hinder accurate user preference modeling, thereby limiting recommendation performance. While existing approaches often emphasize interpretability, they frequently overlook the optimization of user-item affinity in the feature space. To overcome this limitation, the paper proposes DIAURec, which uniquely integrates intent modeling with language modeling to construct a dual intent space—comprising both prototype- and distribution-based representations—driven by collaborative and linguistic signals. Cross-space alignment is achieved through coarse-to-fine granularity matching. The method further introduces an optimization strategy that balances alignment and uniformity, complemented by intra-space and interaction-aware regularization to mitigate representation collapse, thereby enhancing consistency and robustness. Extensive experiments on three public benchmarks demonstrate that DIAURec significantly outperforms fifteen state-of-the-art baselines, confirming its effectiveness and superiority.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Knowledge Representation and Reasoning: Preferences
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

General recommender systems deliver personalized services by learning user and item representations, with the central challenge being how to capture latent user preferences. However, representations derived from sparse interactions often fail to comprehensively characterize user behaviors, thereby limiting recommendation effectiveness. Recent studies attempt to enhance user representations through sophisticated modeling strategies ($e.g.,$ intent or language modeling). Nevertheless, most works primarily concentrate on model interpretability instead of representation optimization. This imbalance has led to limited progress, as representation optimization is crucial for recommendation quality by promoting the affinity between users and their interacted items in the feature space, yet remains largely overlooked. To overcome these limitations, we propose DIAURec, a novel representation learning framework that unifies intent and language modeling for recommendation. DIAURec reconstructs representations based on the prototype and distribution intent spaces formed by collaborative and language signals. Furthermore, we design a comprehensive representation optimization strategy. Specifically, we adopts alignment and uniformity as the primary optimization objectives, and incorporates both coarse- and fine-grained matching to achieve effective alignment across different spaces, thereby enhancing representational consistency. Additionally, we further introduce intra-space and interaction regularization to enhance model robustness and prevent representation collapse in reconstructed space representation. Experiments on three public datasets against fifteen baseline methods show that DIAURec consistently outperforms state-of-the-art baselines, fully validating its effectiveness and superiority.

Problem

• representation optimization
• user preference
• recommendation
• sparse interactions
• intent modeling

Innovation

• representation optimization
• dual-intent space
• alignment and uniformity
• language modeling
• regularization