Abstract:In real-world recommender systems, implicitly collected user feedback, while abundant, often includes noisy false-positive and false-negative interactions. The possible misinterpretations of the user-item interactions pose a significant challenge for traditional graph neural recommenders. These approaches aggregate the users' or items' neighbours based on implicit user-item interactions in order to accurately capture the users' profiles. To account for and model possible noise in the users' interactions in graph neural recommenders, we propose a novel Diffusion Graph Transformer (DiffGT) model for top-k recommendation. Our DiffGT model employs a diffusion process, which includes a forward phase for gradually introducing noise to implicit interactions, followed by a reverse process to iteratively refine the representations of the users' hidden preferences (i.e., a denoising process). In our proposed approach, given the inherent anisotropic structure observed in the user-item interaction graph, we specifically use anisotropic and directional Gaussian noises in the forward diffusion process. Our approach differs from the sole use of isotropic Gaussian noises in existing diffusion models. In the reverse diffusion process, to reverse the effect of noise added earlier and recover the true users' preferences, we integrate a graph transformer architecture with a linear attention module to denoise the noisy user/item embeddings in an effective and efficient manner. In addition, such a reverse diffusion process is further guided by personalised information (e.g., interacted items) to enable the accurate estimation of the users' preferences on items. Our extensive experiments conclusively demonstrate the superiority of our proposed graph diffusion model over ten existing state-of-the-art approaches across three benchmark datasets.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: In real - world recommendation systems, although the implicitly collected user feedback is abundant, it usually contains noisy false positives and false negative interactions. This possible misinterpretation of user - item interactions poses a significant challenge to traditional graph - neural - network - based recommendation systems. These methods aggregate the neighbors of users or items according to implicit user - item interactions to accurately capture user profiles. However, due to the noise in implicit interaction data, traditional methods have difficulty accurately representing users' preferences. To address this problem, the authors propose a new Diffusion Graph Transformer (DiffGT) model for Top - 𝑘 recommendation. Specifically, the model solves the problem in the following ways: 1. **Introducing Directional Gaussian Noise**: Different from the isotropic Gaussian noise used in existing diffusion models, the authors use anisotropic and directional Gaussian noise in the forward diffusion process. This enables the model to better capture the inherent anisotropic and directional structures in the recommendation data. 2. **Denoising in the Reverse Diffusion Process**: In the reverse diffusion process, the authors combine the graph - transformer architecture and the linear - attention module to effectively remove noise and restore the representation of users' hidden preferences. In addition, the reverse diffusion process is guided by personalized information (such as interacted items), thereby more accurately estimating users' item preferences. 3. **Optimized Diffusion Strategy**: The authors not only achieve continuous diffusion at the embedding level but also optimize discrete diffusion on the graph structure. By comparative analysis of these two diffusion strategies, the authors demonstrate the advantages of continuous diffusion in terms of efficiency and effectiveness. ### Formula Summary - **Forward Diffusion Process**: \[ q(x_t|x_{t - 1})=\mathcal{N}(x_t; \sqrt{1 - \beta_t}x_{t - 1}, \beta_tI) \] \[ x_t=\sqrt{1 - \beta_t}x_{t - 1}+\sqrt{\beta_t}\epsilon, \quad \epsilon\sim\mathcal{N}(0, I) \] - **Backward Diffusion Process**: \[ p(x_{t - 1}|x_t, x_0)=\mathcal{N}(x_{t - 1}; \mu_\theta^t(x_t, x_0), \beta_tI) \] \[ \mu_\theta^t(x_t, x_0)=\frac{\sqrt{\bar{\alpha}_{t - 1}}\beta_t}{1 - \bar{\alpha}_t}x_0+\frac{\sqrt{\alpha_t(1 - \bar{\alpha}_{t - 1})}}{1 - \bar{\alpha}_t}x_t \] \[ \beta_t=\frac{1 - \bar{\alpha}_{t - 1}}{1 - \bar{\alpha}_t}\beta_t \] - **Directional Gaussian Noise**: \[ x_t=\sqrt{\bar{\alpha}_t}x_0+\sqrt{1 - \bar{\alpha}_t}\epsilon' \] \[ \epsilon'=\text{sgn}(x_0)\odot|\bar{\epsilon}| \] \[ \bar{\epsilon}=\mu+\sigma\odot\epsilon \] Through these improvements, the DiffGT model can more accurately capture users' preferences when dealing with noisy implicit user - item interactions, thereby improving the performance of the recommendation system.

A Directional Diffusion Graph Transformer for Recommendation

DiffKG: Knowledge Graph Diffusion Model for Recommendation

Graph Diffusion-based Representation Learning for Sequential Recommendation

EDiffuRec: An Enhanced Diffusion Model for Sequential Recommendation

Dynamic Recommendation Based on Graph Diffusion and Ebbinghaus Curve

EDGE-Rec: Efficient and Data-Guided Edge Diffusion For Recommender Systems Graphs

Diffusion Recommender Model

Position-aware Graph Transformer for Recommendation

RecDiff: Diffusion Model for Social Recommendation

Federated Recommender System Based on Diffusion Augmentation and Guided Denoising

Denoising Diffusion Recommender Model

Graph Trend Filtering Networks for Recommendation

TransGNN: Harnessing the Collaborative Power of Transformers and Graph Neural Networks for Recommender Systems

Graph Trend Filtering Networks for Recommendations

Discrete Conditional Diffusion for Reranking in Recommendation

A novel diffusion recommendation algorithm based on multi-scale cnn and residual lstm

DiffNet++: A Neural Influence and Interest Diffusion Network for Social Recommendation

DiffuRec: A Diffusion Model for Sequential Recommendation

Generative Diffusion Models for Sequential Recommendations

Sequential Recommendation with Diffusion Models

Collaborative Filtering Based on Diffusion Models: Unveiling the Potential of High-Order Connectivity