基于扩散模型中噪声重调度的自适应基于物品的协同结构用于生成式推荐
Adaptive Item-based Collaborative Structures via Noise Rescheduling in Diffusion for Generative Recommendation
- Tongji University（同济大学）
- Huawei Technologies Co., Ltd.（华为技术有限公司）
- Fudan University（复旦大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
针对现有生成式推荐模型未整合物品协同信息的缺陷，提出 ANR-DiffRec 框架，通过引入物品共现矩阵与自适应噪声重调度机制，在多基准测试中优于现有最优模型。
AI中文摘要：
离散扩散模型（DDMs）近期被引入推荐系统，将用户历史建模为通过迭代去噪的 token 生成过程。然而，这类方法虽能有效捕捉用户级序列模式，却常无法明确整合基于物品的协同过滤信息——这是精准推荐的关键组成部分。该缺陷体现在两个核心方面：（1）物品表示通常以语义为核心，缺乏扩散训练所需的协同先验；（2）去噪过程采用均匀噪声调度，不加区分地处理所有 token，忽略了物品级自适应结构依赖。为填补这一空白，我们提出 ANR-DiffRec，一个将基于物品的协同结构编码到离散扩散模型以实现生成式推荐的统一框架。首先，我们明确引入物品共现矩阵以指导语义 ID 生成，为离散扩散训练提供结构化协同先验；其次，我们提出基于物品的自适应噪声重调度机制，该机制可根据局部上下文可恢复性和行为感知的物品依赖动态调整去噪权重。具体而言，所提策略联合建模物品内部结构上下文与物品间协同信号，使扩散训练过程中实现结构感知的去噪。在多个基准数据集上开展的大量实验表明，我们的方法始终优于当前最优的生成式推荐模型。代码：this https URL。
英文摘要：
Discrete Diffusion Models (DDMs) have recently been introduced to recommendation systems, modeling user history as a token generation process via iterative denoising. However, while effective at capturing user-level sequential patterns, these methods often fail to explicitly integrate item-based collaborative filtering information, a critical component for accurate recommendation. This deficiency manifests in two key aspects: (1) the item representation is often semantic-focused, lacking collaborative priors for diffusion training; and (2) the denoising process employs a uniform noise schedule, treating all tokens indiscriminately and ignoring item-level adaptive structural dependencies. To bridge this gap, we propose ANR-DiffRec, a unified framework designed to encode item-based collaborative structures into discrete diffusion for generative recommendation. First, we explicitly incorporate an item co-occurrence matrix to guide semantic ID generation, providing a structured collaborative prior for discrete diffusion training. Second, we introduce an item-based adaptive noise rescheduling mechanism that dynamically adjusts denoising weights according to both local contextual recoverability and behavior-aware item dependencies. Specifically, the proposed strategy jointly models intra-item structural context and inter-item collaborative signals, enabling structure-aware denoising during diffusion training. Extensive experiments on multiple benchmarks demonstrate that our method consistently outperforms state-of-the-art generative recommendation models. Code: https://github.com/CalmaQi/ANR-DiffRec.
