Time-Aware Diffusion based on Preference Disentanglement for Generative Recommendation
2026-06-01 • Information Retrieval
Information RetrievalArtificial Intelligence
AI summaryⓘ
The authors study recommendation systems that generate item suggestions using semantic tokens instead of item IDs. They point out that previous methods treat all past items equally over time, which does not reflect how user preferences actually change. To address this, the authors propose TDPM, a new approach that considers both long-term stable preferences and recent short-term interests by adding a time dimension to the diffusion process. Their experiments show that TDPM performs better than existing models on several datasets. They also confirm that including time information in the model is important for better recommendations.
Generative RecommenderSemantic IndexDiffusion ModelsTime-aware DiffusionUser PreferenceNon-stationary DistributionPeriod PreferencePoint PreferenceHR@20NDCG@20
Authors
Bangguo Zhu, Peng Huo, Yuanbo Zhao, Zhicheng Du, Jun Yin, Senzhang Wang
Abstract
Recently, Generative Recommenders (GRs) have emerged as a transformative recommendation paradigm by replacing traditional item IDs with semantic indices (SIDs). Owing to the exceptional generative capabilities of diffusion models, a few pioneering works explore developing GRs with diffusion architectures as the backbone. However, a fatal limitation of existing diffusion-based GRs is that the diffusion process applies uniformly to all items within the historical interactions. In contrast, the user preference is shaped by multifaceted time-evolving factors and thus exhibits a non-stationary distribution in the temporal aspect. To bridge this gap, this study proposes a novel GR framework, named TDPM, by designing the time-aware diffusion on SID tokens. Specifically, TDPM explicitly integrates the impact of time-evolving user preferences into the diffusion process. In detail, the user preference is disentangled into (i) the period preference, which remains consistent over a long time-span, and (ii) the point preference, which is triggered by recent focal events. Extensive experiments on three public real-world datasets demonstrate the significant superiority of TDPM over the state-of-the-art baselines. TDPM achieves average improvements of up to 29.21% and 25.45% in terms of HR@20 and NDCG@20, respectively. The ablation study further underscores the necessity of time-aware token diffusion in diffusion-based GRs.