Graph Contrastive Learning (GCL) has recently garnered significant attention for enhancing recommender systems. Most existing GCL-based methods perturb the raw data graph to generate views, performing contrastive learning across these views to learn generalizable representations. However, most of these methods rely on data- or model-based augmentation techniques that may disrupt interest consistency. In this paper, we propose a novel interest-aware augmentation approach based on diffusion models to address this issue. Specifically, we leverage a conditional diffusion model to generate interest-consistent views by conditioning on node interaction information, ensuring that the generated views align with the interests of the nodes. Based on this augmentation method, we introduce DiffCL, a graph contrastive learning framework for recommendation. Furthermore, we propose an easy-to-hard generation strategy. By progressively adjusting the starting point of the reverse denoising process, this strategy further enhances effective contrastive learning. We evaluate DiffCL on three public real-world datasets, and results indicate that our method outperforms state-of-the-art techniques, demonstrating its effectiveness.

Social recommendation systems leverage both user-user (u-u) social relations and user-item (u-i) collaborative interactions to improve recommendation quality. Despite their effectiveness, existing models often struggle with task-irrelevant information and misalignment between social and collaborative signals and the downstream recommendation task, leading to suboptimal performance. To address these limitations, we propose a novel framework for Effective and Consistent Information Extraction for Social Recommendation (ECSR). Our approach focuses on two key modules: (1) a task-irrelevant information discarding module that filters out noisy signals from both social relations and user-item interactions, and (2) a task-relevant information alignment module that captures both shared and view-specific task-relevant information, ensuring alignment with the recommendation objective. By integrating them into a unified form, our method extracts minimal and sufficient statistics, which significantly enhance the model's ability to predict user preferences. We validate ECSR on three real-world social recommendation datasets, demonstrating that it consistently outperforms state-of-the-art baselines.

Designing effective negative sampling strategies is crucial for training Multi-Modal Recommendation (MMRec) models, as it helps address the issues of sparse user-item interactions and facilitates the learning of high-dimensional modality features. However, most existing methods randomly sample non-interacted items as negative ones, which frequently result in easy negatives. They limit the model's ability to accurately capture user preferences. In this paper, we propose to Generate Difficulty-aware Negative Samples via conditional diffusion for MMRec (denoted as GDNSM). Leveraging the rich semantic and contextual information from multi-modal features, our method generates hard negative samples with varying difficulty levels, tailored to user preferences. They force the model to learn finer-grained distinctions between positive and negative samples, enhancing its ability to inferring user preferences. To avoid unstable training, we design a dynamic difficulty scheduling mechanism that schedules the negative samples from easy to hard for model training, ensuring both stability and effectiveness. Extensive experiments on three real-world datasets demonstrate that the effectiveness of our models.

The integration of social information into recommender systems (RSs) has gained significant popularity for enhancing recommendation performance and user experience. However, this practice introduces substantial privacy risks, particularly concerning the leakage of sensitive social relationships. While prior research has primarily focused on user-level and interaction-level privacy risks, the social relation-level privacy risks remain largely unexplored. To fill this gap, we investigate social privacy risks through membership inference attacks (MIA) and propose a Social relation-level MIA (SMIA) framework. Two key challenges arise: (1) the adversary can only access the recommended item IDs, which provide indirect and limited information about social relationships, and (2) extracting socially relevant preferences from recommendation results is inherently difficult. To tackle the first challenge, we leverage shadow models to transform sparse item IDs into dense features, enabling adversaries to effectively utilize recommendation outputs. For the second challenge, SMIA employs a dual-branch learning approach that disentangles social and behavioral preferences. Therefore, we can extract socially relevant signals from the disentangled preferences.Extensive experiments on real-world datasets demonstrate that both social and general RSs are highly vulnerable to such attacks, highlighting the urgent need for robust privacy protection mechanisms. To defend against these attacks, we introduce a Socially Adversarial Learning (SAL) defense mechanism that selectively obscures sensitive social information in user representations during training, effectively reducing privacy leakage. We further evaluate the effectiveness of our defense and discuss future directions for developing privacy-preserving mechanisms in social RSs.

Graph-based social recommendation improves the prediction accuracy of recommendation by leveraging high-order neighboring information contained in social relations. However, most of them ignore the problem that social relations can be noisy for recommendation. Several studies attempt to tackle this problem by performing social graph denoising, but they suffer from 1) adaptability issues for other graph-based social recommendation models and 2) insufficiency issues for user social representation learning. To address the limitations, we propose a model-agnostic graph denoising module (denoted as MADM) which works as a plug-and-play module to provide refined social structure for base models. Meanwhile, to propel user social representations to be minimal and sufficient for recommendation, MADM further employs mutual information maximization (MIM) between user social representations and the interaction graph and realizes two ways of MIM: contrastive learning and forward predictive learning. We provide theoretical insights and guarantees from the perspectives of Information Theory and Multi-view Learning to explain its rationality. Extensive experiments on three real-world datasets demonstrate the effectiveness of MADM.

Sequential recommendation is a critical part of the flourishing online applications by suggesting appealing items on users’ next interactions, where global dependencies among items have proven to be indispensable for enhancing the quality of item representations toward a better understanding of user dynamic preferences. Existing methods rely on pre-defined graphs with shallow Graph Neural Networks to capture such necessary dependencies due to the constraint of the over-smoothing problem. However, this graph representation learning paradigm makes them difficult to satisfy the original expectation because of noisy graph structures and the limited ability of shallow architectures for modeling high-order relations. In this paper, we propose a novel Graph Diffusion Representation-enhanced Attention Network for sequential recommendation, which explores the construction of deeper networks by utilizing graph diffusion on adaptive graph structures for generating expressive item representations. Specifically, we design an adaptive graph generation strategy via leveraging similarity learning between item embeddings, automatically optimizing the input graph topology under the guidance of downstream recommendation tasks. Afterward, we propose a novel graph diffusion paradigm with robustness to over-smoothing, which enriches the learned item representations with sufficient global dependencies for attention-based sequential modeling. Moreover, extensive experiments demonstrate the effectiveness of our approach over state-of-the-art baselines.

The issue of user cold-start poses a long-standing challenge to recommendation systems, due to the scarce interactions of new users. Recently, meta-learning based studies treat each cold-start user as a user-specific few-shot task and then derive meta-knowledge about fast model adaptation across training users. However, existing solutions mostly do not clearly distinguish the concept of new users and the concept of novel preferences, leading to over-reliance on meta-learning based adaptability to novel patterns. In addition, we also argue that the existing meta-training task construction inherently suffers from the memorization overfitting issue, which inevitably hinders meta-generalization to new users. In response to the aforementioned issues, we propose a preference learning decoupling framework, which is enhanced with meta-augmentation (PDMA), for user cold-start recommendation. To rescue the meta-learning from unnecessary adaptation to common patterns, our framework decouples preference learning for a cold-start user into two complementary aspects: common preference transfer, and novel preference adaptation. To handle the memorization overfitting issue, we further propose to augment meta-training users by injecting attribute-based noises, to achieve mutually-exclusive tasks. Extensive experiments on benchmark datasets demonstrate that our framework achieves superior performance improvements against state-of-the-art methods. We also show that our proposed framework is effective in alleviating memorization overfitting.

Next Point-of-Interest (POI) recommendation is an essential part of the flourishing location-based applications, where the demands of users are not only conditioned by their recent check-in behaviors but also by the critical influence stemming from geographical dependencies among POIs. Existing methods leverage Graph Neural Networks with the aid of pre-defined POI graphs to capture such indispensable correlations for modeling user preferences, assuming that the appropriate geographical dependencies among POIs could be pre-determined. However, the pre-defined graph structures are always far from the optimal graph topology due to noise and adaptability issues, which may decrease the expressivity of learned POI representations as well as the credibility of modeling user preferences. In this paper, we propose a novel Adaptive Graph Representation-enhanced Attention Network (AGRAN) for next POI recommendation, which explores the utilization of graph structure learning to replace the pre-defined static graphs for learning more expressive representations of POIs. In particular, we develop an adaptive POI graph matrix and learn it via similarity learning with POI embeddings, automatically capturing the underlying geographical dependencies for representation learning. Afterward, we incorporate the learned representations of POIs and personalized spatial-temporal information with an extension to the self-attention mechanism for capturing dynamic user preferences. Extensive experiments conducted on two real-world datasets validate the superior performance of our proposed method over state-of-the-art baselines.

Next Point-of-Interest (POI) recommendation plays a critical role in many location-based applications as it provides personalized suggestions on attractive destinations for users. Since users' next movement is highly related to the historical visits, sequential methods such as recurrent neural networks are widely used in this task for modeling check-in behaviors. However, existing methods mainly focus on modeling the sequential regularity of check-in sequences but pay little attention to the intrinsic characteristics of POIs, neglecting the entanglement of the diverse influence stemming from different aspects of POIs. In this paper, we propose a novel Disentangled Representation-enhanced Attention Network (DRAN) for next POI recommendation, which leverages the disentangled representations to explicitly model different aspects and corresponding influence for representing a POI more precisely. Specifically, we first design a propagation rule to learn graph-based disentangled representations by refining two types of POI relation graphs, making full use of the distance-based and transition-based influence for representation learning. Then, we extend the attention architecture to aggregate personalized spatio-temporal information for modeling dynamic user preferences on the next timestamp, while maintaining the different components of disentangled representations independent. Extensive experiments on two real-world datasets demonstrate the superior performance of our model to state-of-the-art approaches. Further studies confirm the effectiveness of DRAN in representation disentanglement.

The task of next Point-of-Interest (POI) recommendation aims at recommending a list of POIs for a user to visit at the next timestamp based on his/her previous interactions, which is valuable for both location-based service providers and users. Recent state-of-the-art studies mainly employ recurrent neural network (RNN) based methods to model user check-in behaviors according to user’s historical check-in sequences. However, most of the existing RNN-based methods merely capture geographical influences depending on physical distance or successive relation among POIs. They are insufficient to capture the high-order complex geographical influences among POI networks, which are essential for estimating user preferences. To address this limitation, we propose a novel Graph-based Spatial Dependency modeling (GSD) module, which focuses on explicitly modeling complex geographical influences by leveraging graph embedding. GSD captures two types of geographical influences, i.e., distance-based and transition-based influences from designed POI semantic graphs. Additionally, we propose a novel Graph-enhanced Spatial-Temporal network (GSTN), which incorporates user spatial and temporal dependencies for next POI recommendation. Specifically, GSTN consists of a Long Short-Term Memory (LSTM) network for user-specific temporal dependencies modeling and GSD for user spatial dependencies learning. Finally, we evaluate the proposed model using three real-world datasets. Extensive experiments demonstrate the effectiveness of GSD in capturing various geographical influences and the improvement of GSTN over state-of-the-art methods.