Contrastive learning has demonstrated promising potential in recommender systems. Existing methods typically construct sparser views by randomly perturbing the original interaction graph, as they have no idea about the authentic user preferences. Owing to the sparse nature of recommendation data, this paradigm can only capture insufficient semantic information. To address the issue, we propose InfoDCL, a novel diffusion-based contrastive learning framework for recommendation. Rather than injecting randomly sampled Gaussian noise, we employ a single-step diffusion process that integrates noise with auxiliary semantic information to generate signals and feed them to the standard diffusion process to generate authentic user preferences as contrastive views. Besides, based on a comprehensive analysis of the mutual influence between generation and preference learning in InfoDCL, we build a collaborative training objective strategy to transform the interference between them into mutual collaboration. Additionally, we employ multiple GCN layers only during inference stage to incorporate higher-order co-occurrence information while maintaining training efficiency. Extensive experiments on five real-world datasets demonstrate that InfoDCL significantly outperforms state-of-the-art methods. Our InfoDCL offers an effective solution for enhancing recommendation performance and suggests a novel paradigm for applying diffusion method in contrastive learning frameworks.

Cross-domain sequential recommendation (CDSR) aims to utilize users' interactions across multiple domains to alleviate the problem of interaction sparsity that is prevalent in web platforms, thereby providing more accurate personalized recommendations. Although current CDSR methods have made some progress, they suffer from two main limitations: (i) assuming uniformly distributed interactions over time; and (ii) neglecting temporal influences during cross-domain transfer. In order to address the above issues, we propose a novel Time-Aware Cross-Domain Sequential Recommendation framework (TA-CDSR ). First, we design a time-sensitive attention which captures user preferences over time by decoupling interaction sequences and time sequences. Second, we propose a time-guided preference generator that can reconstruct the lacking interactions in the target domain by taking the source domain interactions time as guidance information. Finally, we design a multi-scale time windows based domain transfer module, which can dynamically identify the temporal interaction density and thus adaptively assign the weights of cross-domain information. Extensive experiments on three real-world datasets indicate that TA-CDSR achieves competitive time complexity while outperforming other baselines.

Multi-behavior Session Based Recommendations (MBSBRs) have achieved remarkable results due to considering behavioral heterogeneity in sessions. Yet most existing works only consider binary or continuous behavior dependencies and aim to predict the next item under the target behavior, neglecting users' inherent behavior habits, resulting in learning inaccurate intentions. To tackle the above issues, we propose a novelBehaviorHabits Enhanced Intention Learning framework forSessionBasedRecommendation (BHSBR). Specifically, we focus on the next item recommendation and design a global item transition graph to learn the behavior-aware semantic relationships between items, in order to mine the underlying similarity between items beyond the session. In addition, we construct a hypergraph to extract the diverse behavior habits of users and break through the limitations of temporal relationships in the session. Compared to the existing works, our behavior habit learning method learns behavior dependencies at the user level, which could capture the user's more accurate long-term intentions and reduce the impact of noise behaviors. Extensive experiments on three datasets demonstrate that the performance of our proposedBHSBRis superior to SOTA. Further ablation experiments fully illustrate the effectiveness of our various modules.

Session-based recommendation (SBR) aims to predict upcoming user choices based on brief interaction histories. Over the past few years, graph neural networks (GNNs) have become a powerful tool for capturing intricate item relationships and delivering effective recommendations. Existing works use the sequential interactions within all sessions to construct graphs and provide self-supervised signals. Although some progresses have been made, we argue that merely relying on the transitions pattern fail to fully mine the complex information among items and lead to limited item representations. This article introduces a semantic information and intention enhanced SBR paradigm, which is called SISR. Our SISR leverages not only the sequential order of items but also the session intentions and semantic neighbors. Specifically, we begin by constructing a global item transition graph to enhance the GNN-based SBR with insights from items across all sessions. Then, the clustering mechanism is applied to obtain latent semantic prototypes of items and further extract the intention representations of sessions. Finally, we propose two contrastive learning component to distill the self-supervised signals for the item representation learning, so as to alleviate the data-sparsity phenomenon and augment the item recommendation component ratio. Comprehensive testing on three real-world datasets against various leading models highlights the advantages of our SISR paradigm.

Graph-based recommendation systems have excelled in modeling user–item interactions, but most focus solely on positive feedback, overlooking the critical role of negative feedback in capturing comprehensive user preferences. Signed graphs, which incorporate both positive and negative interactions, offer a promising approach but face challenges due to the oversimplification of balance theory and the limitations of conventional graph neural networks (GNNs) in processing negative signals. To address these issues, we propose paired channel enhanced sign-aware graph recommendation (PCSRec), a novel framework that holistically integrates positive and negative feedback. PCSRec introduces a path-enhanced embedding module that leverages a learnable path-encoding matrix to capture indirect structural patterns, overcoming the limitations of balance theory. Additionally, it employs a paired channel filtering mechanism with spectral low-pass and high-pass filters to model similarity from positive feedback and dissimilarity from negative feedback, respectively. A dual-loss optimization strategy, combining contrastive and Bayesian personalized ranking (BPR) losses, further refines discriminative representations. Extensive experiments on four real-world datasets demonstrate that PCSRec outperforms both unsigned and signed graph-based baselines, achieving state-of-the-art recommendation performance. Ablation studies and visualizations confirm the effectiveness of its components in improving embedding quality and recommendation accuracy.

In recent years, social recommendation systems have emerged as a pivotal technology for enhancing recommendation accuracy by leveraging user social homophily and influence. Although many works have been devoted to this area, existing works still struggle to extract the beneficial structural information from social relationships that is beneficial for recommendations and neglect the inherent popularity bias in the social networks, which leads to suboptimal recommendation performances. To address these challenges, we propose a novel framework termed Causal Disentanglement-Enhanced Diffusion Denoising for Social Recommendation (CaDDiSR). This framework first employs causal graphs to disentangle the complexities of social relationships, generating user representations with high-order structures, which are subsequently used as inputs to a diffusion process to effectively denoise social networks and retain social signals beneficial for recommendation tasks. Furthermore, the framework integrates a bidirectional knowledge distillation mechanism, which balances user representations between social and recommendation contexts, thereby facilitating the effective fusion of their respective advantages while simultaneously mitigating noise interference and enhancing overall system performance. Finally, cross-domain contrastive learning is utilized to optimize user and item representations, ensuring consistency in recommendation performance across diverse scenarios. Experimental results on multiple real-world datasets demonstrate that CaDDiSR significantly outperforms existing baseline models, substantiating its superior performance.

Session-based recommendation (SBR) systems have increasingly focused on hypergraph-based approaches due to their potent capability in capturing high-order item relationships. Typically, existing approaches rely on sequential item relations to manually construct fixed hypergraphs. However, this methodology neglects the multiple relations inherent in the original sequences, thereby impeding the hypergraph’s precision in discerning user preferences. Furthermore, the rigidity of fixed hypergraph structures tends to emphasize explicit relationships, ignoring the latent implicit patterns. In light of this, we present a novel Multi-relation enhanced Dynamic HyperGraph (MDHG) learning framework for session-based recommendation, to model intricate and variable item relations. Initially, we establish three distinct relation graphs which capture separate user behavior patterns to extract personalized interest preferences under differentiated intentions. Subsequently, we propose an enhanced dynamic hypergraph paradigm that adaptively generates hypergraph structures based on prior relation graph, thereby reinforcing and unveiling implicit connectivity relations in a layer-aware manner. Finally, to mitigate the noise among diverse relations, we introduce the maximum mutual information auxiliary task and employ the attention mechanism as a cross-relation aggregator. Extensive experiments on various real-world datasets verify the superiority of our MDHG model. Our code is publicly available at https://github.com/Qin-lab-code/MDHG .

Conversational Recommender Systems (CRSs) interact with users through natural language to provide recommendations and generate responses. Due to limited information in conversation, existing works utilize KGs or reviews to improve CRS. Despite achievements, they overlook co-occurrence relations which have shown effectiveness in collaborative filtering systems. In this work, we first propose a novel framework named CoCRS , aiming to incorporate Co-occurrences into the Review-based Conversation Recommendation Systems . In CoCRS, we mine co-occurrences from two aspects: (1) item and entity , (2) user and item . For the first one, we extract entities from redundant review texts by KG and construct a relation-aware item-entity heterogeneous graph. In the second aspect, we analyze review sentiments and construct a sentiment-aware user-item bipartite graph. We encode two graphs to obtain user and entity embeddings. Since users in CRS are anonymous, we generate a virtual similar user representation to match reviews with users. Besides, we capture time-aware preference representation from two-time dimensions. Finally, we generate word-level user representation with word-oriented KG and model user preference by integrating the above representations. Extensive experiments demonstrate that CoCRS outperforms baselines and the cold-start experiment highlights its robustness. The Large Language Model (LLM) experiment illustrates the significant role of co-occurrence relationships in LLM-based CRS. Our code are available at https://github.com/Qin-lab-code/CoCRS .

Pretrained vision-language models (VLMs), e.g., CLIP, demonstrate impressive zero-shot capabilities on downstream tasks. Prior research highlights the crucial role of visual augmentation techniques, like random cropping, in alignment with fine-grained class descriptions generated by large language models (LLMs), significantly enhancing zero-shot performance by incorporating multi-view information. However, the inherent randomness of these augmentations can inevitably introduce background artifacts and cause models to overly focus on local details, compromising global semantic understanding. To address these issues, we propose an Attention-Based Selection (ABS) method from local details to global context, which applies attention-guided cropping in both raw images and feature space, supplement global semantic information through strategic feature selection. Additionally, we introduce a soft matching technique to effectively filter LLM descriptions for better alignment. ABS achieves state-of-the-art performance on out-of-distribution generalization and zero-shot classification tasks. Notably, ABS is training-free and even rivals few-shot and test-time adaptation methods.

Sequential recommendation (SR) is widely deployed in e-commerce platforms, streaming services, etc., revealing significant potential to enhance user experience. The core of SR lies in exploring the sequential relationships in historical user-item interactions. However, existing methods often overlook two critical factors: irregular user interests between interactions and highly uneven item distributions over time. The former factor implies that actual user preferences are not always continuous, and long-term historical interactions may not be relevant to current purchasing behavior. Therefore, relying only on these historical interactions for recommendations may result in a lack of user interest at the target time. The latter factor, characterized by peaks and valleys in interaction frequency, may result from seasonal trends, special events, or promotions. These externally driven distributions may not align with individual user interests, leading to inaccurate recommendations. To address these deficiencies, we propose TGODE to both enhance and capture the long-term historical interactions. Specifically, we first construct the user time graph and item evolution graph, which utilize user personalized preferences and global item distribution information, respectively. To tackle the temporal sparsity caused by irregular user interactions, we design a time-guided diffusion generator to automatically obtain an augmented time-aware user graph. Additionally, we devise a user interest truncation factor to efficiently identify sparse time intervals and achieve balanced preference inference. After that, the augmented user graph and item graph are fed into a generalized graph neural ordinary differential equation (ODE) to align with the evolution of user preferences and item distributions. This allows two patterns of information evolution to be matched over time. Experimental results demonstrate that TGODE outperforms baseline methods across five datasets, with improvements ranging from 10% to 46%. The code is available at https://github.com/Qin-lab-code/TGODE.

Large Language Models (LLMs) have attracted significant attention in recommender systems for their excellent world knowledge capabilities. However, existing methods that rely on Euclidean space struggle to capture the rich hierarchical information inherent in textual and semantic data, which is essential for capturing user preferences. The geometric properties of hyperbolic space offer a promising solution to address this issue. Nevertheless, integrating LLMs-based methods with hyperbolic space to effectively extract and incorporate diverse hierarchical information is non-trivial. To this end, we propose a model-agnostic framework, named HyperLLM, which extracts and integrates hierarchical information from both structural and semantic perspectives. Structurally, HyperLLM uses LLMs to generate multi-level classification tags with hierarchical parent-child relationships for each item. Then, tag-item and user-item interactions are jointly learned and aligned through contrastive learning, thereby providing the model with clear hierarchical information. Semantically, HyperLLM introduces a novel meta-optimized strategy to extract hierarchical information from semantic embeddings and bridge the gap between the semantic and collaborative spaces for seamless integration. Extensive experiments show that HyperLLM significantly outperforms recommender systems based on hyperbolic space and LLMs, achieving performance improvements of over 40%. Furthermore, HyperLLM not only improves recommender performance but also enhances training stability, highlighting the critical role of hierarchical information in recommender systems.

As a typical spatio-temporal series prediction task, traffic flow prediction has found wide application in Intelligent Transportation Systems (ITS). Despite some progress, several unresolved issues persist. Many existing works calculate the dependencies between nodes based on stable long-term traffic data. However, the short-term dependencies are dynamically changing over time, and neglecting them would cause a decrease in predictive performance. In this paper, we propose a novel Dynamic Graph Convolution and Spatio-Temporal Self-Attention (DGSTA) network for traffic flow prediction. Specifically, considering the large amount of short-term and the dynamic dependencies between nodes, we design a new dynamic graph convolution module, which generates adjacency matrices for each time step in a day to dynamically capture the changing short-term dependencies. Additionally, we utilize a multi-head spatio-temporal self-attention module to respectively extract static spatial and temporal correlations between nodes. Furthermore, we design a sequential embedding to explicitly model the long-term correlation between nodes. Extensive experiments conducted on three real-world datasets demonstrate that DGSTA exhibits high competitiveness. The code and data are available at https://github.com/lzmmm30/DGSTA.

Social recommendation (SocialRS), which utilizes user social information to improve recommendation performance, has received increasing attention. Graph neural networks (GNNs) facilitate the integration of both user preference and social features in SocialRS. However, existing techniques face two challenges: 1) the inherent sparse supervision signals and noise issues in real-world social networks; 2) current social recommendation methods suffer from the neglect of user preference and social attribute heterogeneity, which hinders the extraction of preference-related information from social networks. Taking inspiration from social enhancement and contrastive learning methods, we propose a social recommendation model DSVC based on dual social view contrastive learning. Specifically, in response to the first challenge, our model derives the consistency factors of users in different augmented social views, which are used to highlight noise-resistant users and jettison preference-independent social relationships in social views. To address the second challenge, we adopt probability vectors generated from consistency factors. These vectors guide the cross-view augmentation process of the interaction graph, which helps supplement social self-supervised signals and effectively avoid noise retained due to indiscriminate augmentation. The baseline model comparison experiment, ablation experiment, parameter adjustment experiment and robustness experiment conducted on three different real-world datasets consistently validated the effectiveness of our model in improving recommendation performance.

The best arm identification problem in multi-armed bandit model has been widely applied into many practical applications, such as spectrum sensing, online advertising, and cloud computing. Although lots of works have been devoted into this area, most of them do not consider the cost of pulling actions, i.e., a player has to pay some cost when she pulls an arm. Motivated by this, we study a ratio-based best arm identification problem, where each arm is associated with a random reward as well as a random cost. For any \(\delta\in(0,1)\) , with probability at least \(1-\delta\) , the player aims to find the arm with the largest ratio of expected reward to expected cost using as few samplings as possible. Specifically, we consider two settings: (1) the precise setting, i.e., identifying the precise optimal one; (2) the Probably Approximate Correct (PAC) setting, which identifies the \(\epsilon\) -optimal one. For the precise setting, we design the elimination-type algorithms and provide a fundamental lower bound which asymptotically matches the upper bound, while in the PAC setting, an UCB-type algorithm which amed \(\epsilon\) -RCB algorithm is proposed. We show that for all algorithms, the sample complexities, i.e., the pulling times for all arms, grow logarithmically as \(\frac{1}{\delta}\) increases. Moreover, compared to existing works, the running of our algorithms is independent of the arm-related parameters, which is more practical. Finally, we validate our theoretical results through numerical experiments.

Hyperbolic space based collaborative filtering has emerged as a popular topic in recommender systems. Compared to the euclidean space, hyperbolic space is more suitable to the tree-like structures in the user-item interactions and can achieve better recommender performance. Although some works have been devoted to this popular topic and made some progresses, they use tangent space as an approximation of hyperbolic space to implement model. Despite the effectiveness, such methods fail to fully exploit the advantages of hyperbolic space and still suffer from the data sparsity issue, which severely limits the recommender performance. To tackle these problems, we refer to the self-supervised learning technique and novelly propose aHyperbolic GraphContrastiveLearning (HyperCL) framework. Specifically, our framework encodes the augmentation views from both the tangent space and the hyperbolic space, and construct the contrast pairs based on their corresponding learned node representations. Our model not only leverages the geometric advantages of both sides but also achieves seamless information transmission between the two spaces. Extensive experimental results on public benchmark datasets demonstrate that our model is highly competitive and outperforms leading baselines by considerable margins. Further experiments validate the robustness and the superiority of contrastive learning paradigm.

The emerging federated learning enables distributed autonomous vehicles to train equipped deep learning models collaboratively without exposing their raw data, providing great potential for utilizing explosively growing autonomous driving data. However, considering the complicated traffic environments and driving scenarios, deploying federated learning for autonomous vehicles is inevitably challenged by non-independent and identically distributed (Non-IID) data of vehicles, which may lead to failed convergence and low training accuracy. In this paper, we propose a novel hierarchically Federated Region-learning framework of Autonomous Vehicles (FedRAV) that adaptively divides a large area containing vehicles into sub-regions based on the defined region-wise distance, and achieves personalized vehicular models and regional models. Specifically, the architecture employs a designated hypernetwork to learn personalized mask vectors per vehicle used in the linear combination of models shared by vehicles in the same region. This approach ensures that the updated vehicular model adopts the beneficial models while discarding the unprofitable ones. We validate our FedRAV framework against existing federated learning algorithms on four real-world autonomous driving datasets in various heterogeneous settings. Extensive experiment results demonstrate that FedRAV framework achieves superior performance than the state-of-the-art algorithms, and improves the accuracy by 9.36%. The source code of FedRAV is available at:https://github.com/yjzhai-cs/FedRAV.

The emerging federated learning enables distributed autonomous vehicles to train equipped deep learning models collaboratively without exposing their raw data, providing great potential for utilizing explosively growing autonomous driving data. However, considering the complicated traffic environments and driving scenarios, deploying federated learning for autonomous vehicles is inevitably challenged by non-independent and identically distributed (Non-IID) data of vehicles, which may lead to failed convergence and low training accuracy. In this paper, we propose a novel hierarchically Federated Region-learning framework of Autonomous Vehicles (FedRAV), a two-stage framework, which adaptively divides a large area containing vehicles into sub-regions based on the defined region-wise distance, and achieves personalized vehicular models and regional models. This approach ensures that the personalized vehicular model adopts the beneficial models while discarding the unprofitable ones. We validate our FedRAV framework against existing federated learning algorithms on three real-world autonomous driving datasets in various heterogeneous settings. The experiment results demonstrate that our framework outperforms those known algorithms, and improves the accuracy by at least 3.69%. The source code of FedRAV is available at: https://github.com/yjzhai-cs/FedRAV.

Vehicle speed prediction is crucial for intelligent transportation systems, promoting more reliable autonomous driving by accurately predicting future vehicle conditions. Due to variations in drivers’ driving styles and vehicle types, speed predictions for different target vehicles may significantly differ. Existing methods may not realize personalized vehicle speed prediction while protecting drivers’ data privacy. We propose a Federated learning framework with Personalized Aggregation Weights (FedPAW) to overcome these challenges. This method captures client-specific information by measuring the weighted mean squared error between the parameters of local models and global models. The server sends tailored aggregated models to clients instead of a single global model, without incurring additional computational and communication overhead for clients. To evaluate the effectiveness of FedPAW, we collected driving data in urban scenarios using the autonomous driving simulator CARLA, employing an LSTM-based Seq2Seq model with a multi-head attention mechanism to predict the future speed of target vehicles. The results demonstrate that our proposed FedPAW ranks lowest in prediction error within the time horizon of 10 seconds, with a 0.8% reduction in test MAE, compared to eleven representative benchmark baselines.

Diversified recommendation systems have gained increasing popularity in recent years. Nowadays, the emerged graph neural networks (GNNs) have been used to improve the diversity performance. Although some progresses have been made, existing works purely focus on the user–item interactions and overlook the category information, which limits the capability to capture complex diversification among users or items and leads to poor performance. In this article, our target is to integrate full category information into user and item embeddings. To this end, we propose a diversified GNN-based recommendation systems diversified graph recommendation with contrastive learning (DCL). Specifically, we design three key components in our model: 1) the user–item interaction with category-related sampling enhances the interaction of unpopular items; 2) contrastive learning between users and categories shortens the distance of representations between users and their uninteracted categories; and 3) contrastive learning between items and categories diverges the distance of representations between items and their corresponding categories. By applying these three modules, we build a multitask training framework to achieve a balance between accuracy and diversity. Experiments on real-world datasets show that our proposed DCL achieves optimal diversity while paying a little price for accuracy.

The best arm identification problem in multi-armed bandit model has been widely applied into many practical applications, such as spectrum sensing, online advertising, and cloud computing. Although lots of works have been devoted into this area, most of them do not consider the cost of pulling actions, i.e., a player has to pay some cost when she pulls an arm. Motivated by this, we study a ratio-based best arm identification problem, where each arm is associated with a random reward as well as a random cost. For any δ ∈ (0, 1), with probability at least 1-δ, the player aims to find the optimal arm with the largest ratio of expected reward to expected cost using as few samplings as possible. To solve this problem, we propose three algorithms: 1) a genie-aided algorithm GA; 2) the successive elimination algorithm with unknown gaps SEUG; 3) the successive elimination algorithm with unknown gaps and variance information SEUG-V, where gaps denote the differences between the optimal arm and the suboptimal arms. We show that for all three algorithms, the sample complexities, i.e., the pulling times for all arms, grow logarithmically as 1/δ increases. Moreover, compared to existing works, the running of our elimination-type algorithms is independent of the arm-related parameters, which is more practical. In addition, we also provide a fundamental lower bound for sample complexities of any algorithms under Bernoulli distributions, and show that the sample complexities of the proposed three algorithms match that of the lower bound in the sense of log 1/δ. Finally, we validate our theoretical results through numerical experiments.

Cache-enabled wireless networks have gained increasing popularity in recent years for facilitating the delivery of content. Most existing works assumed that the network is static without the change of users and content, which is not the case in the real world. In this paper, we study the content delivery in cache-enabled wireless networks, where the network evolves in terms of users and content. We adopt the affiliation network model to reveal the network evolution process, where new users and content are added into the network randomly and connected with existing nodes via the preferential attachment. Users with strong social relations tend to request the same content objects, thus content popularity is sharply concentrated. Therefore, caching technique is more effective. To maximize the content delivery rate, we formulate the optimization problem of caching replication jointly with routing strategy. We show that network evolution can greatly improve the content delivery rate. In particular, when per-user cache capacity is in the order of Θ(1), per-user delivery rate can achieve constant, i.e., the network can scale. Finally, theoretical results are validated based on a real-world data set from Facebook.

The widespread of smart phones has brought new social-aware features to wireless networks. Users prefer to forward traffic to their social contacts in social-aware networks, which is highly different from the traditional uniform traffic pattern. To this end, we analyze a social-aware wireless network, which is modeled by the social contacts between nodes, and explore its impact on network throughput capacity. We present that it can refine social contacts, thus improving the network capacity. Moreover, directional antennas are applied to further enhance the capacity performance, which can reduce the interference brought by other simultaneous communications. We derive the throughput capacity for social-aware networks with single-beam and multi-beam directional antenna, respectively. For single-beam directional antenna, we prove that when wireless networks are dominated by traffics of short-distance social contact, the throughput capacity can be promoted from order of (θ(1/θ2n log n)1/2) to Θ(1/θ2log n). In addition, when the beamwidth θ is at order of Ω(1/(log n)1/2), the throughput capacity is a constant. Thus, the wireless network is scalable. For multi-beam antennas, we prove that when sidelobe gain Gsis at order of o(Gm), where Gmis main lobe gain, compared with the single-beam case, the network capacity can achieve the order of θ(1/θ2log n).

When social behaviors are considered into ad hoc wireless networks, the users prefer to forward the traffic from their social contacts. In this way, the network capacity is highly related with the social characteristics. To this end, we apply a social contact based wireless network model and explore its throughput capacity. The wireless network capacity will get improved when the contacts are located nearby the source node. Directional antennas are applied to further enhance the capacity performance, which can reduce the interference brought by other simultaneous communications. Our results demonstrate that when the wireless communications are dominated by the short-distance social contact traffics, the throughput capacity can be greatly improved. Meanwhile, when the directional antenna beamwidth is at the order of Ω( 1 √log ), the throughput capacity is a constant. The wireless network with social contacts is scalable.