Predicting diseases solely from patient-side information, such as demographics and self-reported symptoms, has attracted significant research attention due to its potential to enhance patient awareness, facilitate early healthcare engagement, and improve healthcare system efficiency. However, existing approaches encounter critical challenges, including imbalanced disease distributions and a lack of interpretability, resulting in biased or unreliable predictions. To address these issues, we propose the Knowledge graph-enhanced, Prototype-aware, and Interpretable (KPI) framework. KPI systematically integrates structured and trusted medical knowledge into a unified disease knowledge graph, constructs clinically meaningful disease prototypes, and employs contrastive learning to enhance predictive accuracy, which is particularly important for long-tailed diseases. Additionally, KPI utilizes large language models (LLMs) to generate patient-specific, medically relevant explanations, thereby improving interpretability and reliability. Extensive experiments on real-world datasets demonstrate that KPI outperforms state-of-the-art methods in predictive accuracy and provides clinically valid explanations that closely align with patient narratives, highlighting its practical value for patient-centered healthcare delivery.

Relational databases (RDBs) underpin the majority of global data management systems, where information is structured into multiple interdependent tables. In social media platforms, for instance, massive user-generated data are organized across related tables such as users, posts, comments, and interactions, enabling large-scale analysis and predictive modeling of social behaviors. To effectively use the knowledge within RDBs for predictive tasks, recent advances leverage graph representation learning to capture complex inter-table relations as multi-hop dependencies. Despite achieving state-of-the-art performance, these methods remain hindered by the prohibitive storage overhead and excessive training time, due to the massive scale of database and the computational burden of intensive message passing across interconnected tables. To alleviate these concerns, we propose and study the problem of Relational Database Distillation (RDD). Specifically, we aim to distill large-scale RDBs into compact heterogeneous graphs while retaining the predictive power (i.e., utility) required for training graph-based models. Multi-modal column information is preserved through node features, and primary–foreign key relations are encoded via heterogeneous edges, thereby maintaining both data fidelity and relational structure. To ensure adaptability across diverse downstream tasks without engaging the traditional, inefficient bi-level distillation framework, we further design a kernel ridge regression-guided objective with pseudo-labels, which produces quality features for the distilled graph. Extensive experiments on multiple real-world RDBs demonstrate that our solution substantially reduces the data size while maintaining competitive performances on classification and regression tasks, creating an effective pathway for scalable learning with RDBs.

Temporal Graph Neural Networks (TGNNs) are increasingly applied in dynamic scenarios, however, their limited explainability hinders their adoption in high-stakes domains. Existing methods tend to conflate causality with temporal proximity, leading to ambiguous explanations that mix impactful and irrelevant events. Moreover, they lack counterfactual reasoning to assess whether altering specific temporal events would change TGNN predictions. To overcome these challenges, we propose CTM-Explainer, which identifies critical temporal dependencies through iterative “what-if” perturbation analysis. To the best of our knowledge, this is the first post-hoc counterfactual explanation framework for TGNN. It enables precise attribution of how specific timestamped events influence TGNN predictions. By embedding causal analysis into a reinforcement learning framework, CTM-Explainer constructs Counterfactual Temporal Motifs (CTMs) that are causally grounded in model outcome shifts via interventional probability estimation. This design eliminates temporally correlated but non-essential events, while preserving those with verified causal influence. Extensive experiments on real-world and synthetic datasets confirm that CTM-Explainer generates more faithful and concise explanations than existing methods, at significantly lower computational cost.

With growing concerns over user privacy, federated recommendation (FedRec) has emerged as a mainstream solution for personalized recommendation services. FedRec trains user-private parameters on local clients while collaboratively updating global parameters on a centralized server. However, despite advances in optimizing these local and global parameters, existing methods overlook two key challenges: tradeoff training and distribution discrepancy . Tradeoff training balances timely local updates with diverse global parameters, limiting the model’s learning ability. Distribution discrepancy arises from the divergence between locally trained global parameters and those aggregated by the server, corrupting inference performance. To fill in the gap, we propose FedSC , a principled federated recommendation framework that boosts FedRec’s training and inference processes with minimal yet nontrivial efforts. During training, FedSC employs a stochastic aggregation strategy where all users participate in every round, while only a random subset is selected for aggregation, preserving the diversity of global parameters and ensuring timely local updates. During inference, FedSC makes recommendations with a consistency inference mechanism that uses the most recent locally trained global parameters of each user to improve the model’s understanding of user preferences. Extensive experiments on multiple benchmark datasets demonstrate the superiority of FedSC, achieving up to a 20% improvement in most evaluation scenarios.

Contrastive Language-Image Pre-training (CLIP) has emerged as an effective framework for multi-modal representation learning, achieving notable success in diverse tasks such as medical image analysis. CLIP's growing prominence in medical image applications is restricted by its significant computational demands, creating implementation challenges in resource-constrained clinical environments. While knowledge distillation offers an effective approach for model compression with preserved accuracy, existing methods suffer from two fundamental limitations. Firstly, existing methods focus on learning better information from single models while ignoring the fact that student models can generalize well under the guidance of multiple teachers. Secondly, they overlook the complementary information in the CLIP model where the text encoder and image encoder can be leveraged as heterogeneous information to teach one single modality. To tackle these challenges, we propose an Adaptive Heterogeneous Multi-teacher Knowledge Distillation (AdaHet-MKD) framework for effective knowledge transfer across heterogeneous text-image models and among multiple teacher models. The key innovations include: (i) adaptively determining the contribution of each teacher model to specific instances, thereby generating integrated soft logits, and (ii) enabling the student model to operate independently of the teacher model's architecture, which enhances flexibility in teacher-student pairings. Experimental evaluations on publicly available medical datasets demonstrate that our approach has achieved the state-of-the-art performance compared to baselines.

Querying desired data from the key-value store on a cloud server is a prevalent scenario. Client queries might include sensitive information that the client prefers to keep confidential from the server. This occasion resembles the Keyword Private Information Retrieval (KPIR). Prior works on keyword PIR consider that there are no duplicated key-value pairs in the store, i.e., each key only occurs once with only a single value attached. This is one of the cases in practical applications. However, there is also a typical case where a key may occur multiple times with different values. Straightly applying the existing keyword PIR to this case doesn't work and may finally obtain a false query result. We are the first to extend the setting that keys in the store may appear with different values multiple times. To solve this problem, we propose FEDPIR, a fast single-server keyword PIR protocol that supports querying a large-scale key-value store with multiple values per key. FEDPIR uses a novel encoding and decoding strategy combined with a high-throughput linear homomorphic encryption to improve performance significantly. Our extensive experiments on different store configurations show that our FEDPIR achieves 1.2-65.6x lower query latency and 1.5-37.9x lower cost monetarily compared with the baseline methods.

Sequential recommender systems (SRSs) excel in capturing users' dynamic interests, thus playing a key role in various industrial applications. The popularity of SRSs has also driven emerging research on their security aspects, where data poisoning attack for targeted item promotion is a typical example. Existing attack mechanisms primarily focus on increasing the ranks of target items in the recommendation list by injecting carefully crafted interactions (i.e., poisoning sequences), which comes at the cost of demoting users' real preferences. Consequently, noticeable recommendation accuracy drops are observed, restricting the stealthiness of the attack. Additionally, the generated poisoning sequences are prone to substantial repetition of target items, which is a result of the unitary objective of boosting their overall exposure and lack of effective diversity regularizations. Such homogeneity not only compromises the authenticity of these sequences, but also limits the attack effectiveness, as it ignores the opportunity to establish sequential dependencies between the target and many more items in the SRS. To address the issues outlined, we propose a Diversity-aware Dual-promotion Sequential Poisoning attack method named DDSP for SRSs. Specifically, by theoretically revealing the conflict between recommendation and existing attack objectives, we design a revamped attack objective that promotes the target item while maintaining the relevance of preferred items in a user's ranking list. We further develop a diversity-aware, auto-regressive poisoning sequence generator, where a re-ranking method is in place to sequentially pick the optimal items by integrating diversity constraints. By attacking two representative SRSs on three real-world datasets, comprehensive experimental results demonstrate that DDSP outperforms state-of-the-art attack methods in attack effectiveness. Moreover, DDSP achieves the strongest stealthiness with its lowest impact on recommendation accuracy.

The rapid spread of rumors on social media has posed significant challenges to maintaining public trust and information integrity.Since an information cascade process is essentially a propagation tree, recent rumor detection models leverage graph neural networks to additionally capture information propagation patterns, thus outperforming text-only solutions.Given the variations in topics and social impact of the root node, different source information naturally has distinct outreach capabilities, resulting in different heights of propagation trees.This variation, however, impedes the data-driven design of existing graph-based rumor detectors.Given a shallow propagation tree with limited interactions, it is unlikely for graph-based approaches to capture sufficient cascading patterns, questioning their ability to handle less popular news or early detection needs.In contrast, a deep propagation tree is prone to noisy user responses, and this can in turn obfuscate the predictions.In this paper, we propose a novel Epidemiology-informed Network (EIN) that integrates epidemiological knowledge to enhance performance by overcoming data-driven methods' sensitivity to data quality.Meanwhile, to adapt epidemiology theory to rumor detection, it is expected that each user's stance toward the source information will be annotated.To bypass the costly and time-consuming human labeling process, we take advantage of large language models to generate stance labels, facilitating optimization objectives for learning epidemiology-informed representations.Our experimental results demonstrate that the proposed EIN not only outperforms state-of-the-art methods on real-world datasets but also exhibits enhanced robustness across varying tree depths.We release the code at https://github.com/WeiJiang01/EIN.

Network motifs provide a deep insight into the network functional abilities, and have proven useful in various practical applications. Existing studies reveal that different definitions of motifs may be needed for different temporal networks. In this study, we focus on a class of temporal networks such that the nodes and edges keep fixed, but the edge labels vary regularly with timestamps. First, we propose a proper definition of temporal motifs, which appear continuously within sufficiently large time intervals, to properly reinterpret the recurrent and statistically significant nature of motifs in temporal networks. Second, we develop a low polynomial time solution to find temporal motifs for all possible time intervals with the top to bottom and right to left scheme, based on the analyses of the properties for temporal motifs. Third, we develop a theoretically faster incremental solution to efficiently find temporal motifs to support continuously updates of temporal networks, by identifying unaffected time intervals and unnecessary edges. Finally, we have conducted extensive experiments to verify the efficiency and usefulness of our static and incremental solutions.

Graph Neural Networks (GNNs) hold promise in various application domains, but their limited explainability hinders widespread adoption, impacting customer satisfaction and loyalty. This issue intensifies when addressing diverse explanation needs of different user groups. Current GNN explanation models focus on a single objective, neglecting varied and potential conflicting user requirements, resulting in suboptimal outcomes. Moreover, existing models prioritize explanation objectives during multi-objective explanations, disrupting the intrinsic hierarchical structures and distant relationships within the graphs, further diminishing their effectiveness. To tackle these challenges, this paper introduces a novel multi-objective explanatory framework with hierarchical structure attribution for GNNs, termed HM-Explainer. This framework constructs a multi-objective explanation generation module based on Pareto theory to balance different and potentially conflicting explanatory objectives. Additionally, to embed hierarchical information into explanations, HM-Explainer designs node-level and cluster-level attribution modules to analyze the impact of input data on GNN decisions hierarchically. Furthermore, a self-attention mechanism is integrated into the node-level attribution module to account for the influence of distant neighbors. Ultimately, the efficacy of HM-Explainer is validated across multiple datasets for different GNN models through experimentation.

Since the creation of the Web, recommender systems (RSs) have been an indispensable personalization mechanism in information filtering. Most state-of-the-art RSs primarily depend on categorical features such as user and item IDs, and use embedding vectors to encode their information for accurate recommendations, resulting in an excessively large embedding table owing to the immense feature corpus. To prevent the heavily parameterized embedding table from harming RSs’ scalability, both academia and industry have seen increasing efforts compressing RS embeddings, and this trend is further amplified by the recent uptake in edge computing for online services. However, despite the prosperity of existing lightweight embedding-based RSs (LERSs), a strong diversity is seen in the evaluation protocols adopted across publications, resulting in obstacles when relating the reported performance of those LERSs to their real-world usability. On the other hand, among the two fundamental recommendation tasks, namely traditional collaborative filtering and content-based recommendation, despite their common goal of achieving lightweight embeddings, the outgoing LERSs are designed and evaluated with a straightforward “either-or” choice between the two tasks. Consequently, the lack of discussions on a method’s cross-task transferability will likely hinder the development of unified, more scalable solutions for production environments. Motivated by these unresolved issues, this study aims to systematically investigate existing LERSs’ performance, efficiency, and cross-task transferability via a thorough benchmarking process. To create a generic, task-independent baseline, we propose an efficient embedding compression approach based on magnitude pruning, which is proven to be an easy-to-deploy yet highly competitive baseline that outperforms various complex LERSs. Our study reveals the distinct performance of different LERSs across the two recommendation tasks, shedding light on their effectiveness and generalizability under different settings. Furthermore, to account for edge-based recommendation—an increasingly popular use case of LERSs, we have also deployed and tested all LERSs on a Raspberry Pi 4, where their efficiency bottleneck is exposed compared with GPU-based deployment. Finally, we conclude this article with critical summaries on the performance comparison, suggestions on model selection based on task objectives, and underexplored challenges around the applicability of existing LERSs for future research. To encourage and support future LERS research, we publish all source codes and data, checkpoints, and documentation at https://github.com/chenxing1999/recsys-benchmark .

Urban flow prediction is a spatio-temporal modelling task that estimates the throughput of transportation services like buses, taxis, and ride-sharing, where data-driven models have become the most popular solution in the past decade. Meanwhile, the implicitly learned mapping between historical observations to the prediction targets tend to over-simplify the dynamics of real-world urban flows, leading to suboptimal predictions. Some recent spatio-temporal prediction solutions bring remedies with the notion of physics-guided machine learning (PGML), which describes spatio-temporal data with nuanced and principled physics laws, thus enhancing both the prediction accuracy and interpretability. However, these spatio-temporal PGML methods are built upon a strong assumption that the observed data fully conforms to the differential equations that define the physical system, which can quickly become ill-posed in urban flow prediction tasks. The observed urban flow data, especially when sliced into time-dependent snapshots to facilitate predictions, is typically incomplete and sparse, and prone to inherent noise incurred in the collection process (e.g., uncalibrated traffic sensors). As a result, such physical inconsistency between the data and PGML model significantly limits the predictive power and robustness of the solution. Moreover, due to the interval-based predictions and intermittent nature of data filing (e.g., one record per 30 minutes) in many transportation services, the instantaneous dynamics of urban flows can hardly be captured, rendering differential equation-based continuous modelling a loose fit for this setting. To overcome the challenges, we develop a discretized physics-guided network (PN), and propose a data-aware framework Physics-guided Active Sample Reweighting (P-GASR) to enhance PN. Technically, P-GASR incorporates an active sample reweighting pipeline, which not only minimizes the model uncertainty of PN to enhance robustness, but also prioritizes data samples that exhibit higher physical compliance to reinforce their contribution to PN training. Experimental results in four real-world datasets demonstrate that our method achieves state-of-the-art performance with a demonstrable improvement in robustness. The code is released at https://github.com/WeiJiang01/P-GASR.

The drug recommendation task aims to predict safe and effective drug prescriptions based on the patients' historical electronic health records (EHRs). However, existing drug recommendation models generally have two limitations. First, they neglect the inherent characteristics of multiple views existing in patients' clinical data (e.g., diagnoses and procedures), leading to fragmented and inconsistent patient representations. Second, they do not fully exploit drug label information. Most models do not explicitly establish a mapping relationship between drug labels and patients' historical visits. To address these two problems, we proposed a label-aware multi-view drug recommendation model named LAMRec. In particular, LAMRec uses a cross-attention module to fuse information from the diagnosis and procedure views, and increases the mutual information of patient multi-view representations through multi-view contrastive loss; the label-wise attention mechanism fully explores drug label information by constructing an adaptive mapping of drug-visit to generate personalized representations that are aware of the drug-related visit information. Experiments on three real world medical datasets demonstrated the superiority of LAMRec, with a relative reduction of 5.25% in DDI compared to the optimal baseline, a relative improvement of 4.20% in Jaccard similarity scores, and a relative improvement of 3.10% in F1 scores. We released the code online at: https://github.com/Tyunsen/LAMRec.

Food recommendation systems play a pivotal role in shaping dietary salubrity and fostering sustainable lifestyles by recommending recipes and foodstuffs that align with user preferences. Metadata information of a recipe, encompassing multi-modal descriptions, constituent ingredients, and health-related attributes, can furnish a more holistic perspective on the recipe's profile, thereby augmenting recommendation performance. However, existing state-of-the-art methods often overlook the inherent interdependencies between modalities, ingredients, and health factors, leaving the health information pertaining to recipe characteristics underexploited. Notably, our preliminary investigation on two datasets unveiled that the semantic divergence between health-related knowledge and collaborative filtering signals is more pronounced in comparison to other metadata information, thereby potentially impeding the efficacy of food recommendation systems. To address these limitations, we propose HealthRec, a novel multi-modal food recommendation framework with health-aware knowledge distillation. HealthRec employs a global graph representation learning module to capture high-order dependencies across diverse food-related relations, enriching the representations. Subsequently, a co-attention network is leveraged to capture local, recipe-level knowledge transfer between modality-related and ingredient-related embeddings. Additionally, we exploit external supervision signals derived from WHO recommendations, utilizing knowledge distillation during the training phase to transfer local health-aware knowledge into global collaborative embeddings. Extensive experimentation on real-world datasets demonstrates HealthRec's superiority compared to current state-of-the-art recommendation baselines, highlighting its effectiveness in modeling health-aware food recommendations.

Owing to the nature of privacy protection, feder-ated recommender systems (FedRecs) have garnered increasing interest in the realm of on-device recommender systems. However, most existing FedRecs only allow participating clients to collaboratively train a recommendation model of the same public parameter size. Training a model of the same size for all clients can lead to suboptimal performance since clients possess varying resources. For example, clients with limited training data may prefer to train a smaller recommendation model to avoid excessive data consumption, while clients with sufficient data would benefit from a larger model to achieve higher recommendation accuracy. To address the above challenge, this paper introduces HeteFedRec, a novel FedRec framework that enables the assignment of personalized model sizes to partici-pants. Specifically, we present a heterogeneous recommendation model aggregation strategy, including a unified dual-task learning mechanism and a dimensional decorrelation regularization, to allow knowledge aggregation among recommender models of different sizes. Additionally, a relation-based ensemble knowledge distillation method is proposed to effectively distil knowledge from heterogeneous item embeddings. Extensive experiments conducted on three real-world recommendation datasets demonstrate the effectiveness and efficiency of HeteFedRec in training federated recommender systems under heterogeneous settings.

Graph neural networks (GNNs) are currently one of the most performant and versatile collaborative filtering methods. Meanwhile, like in traditional collaborative filtering, owing to the use of an embedding table to represent each user/item entity as a distinct vector, GNN-based recommenders have inherited its long-standing defect of parameter inefficiency. As a common practice for scalable embeddings, parameter sharing enables the use of fewer embedding vectors (which we term meta-embeddings), where each entity is represented by a unique combination of meta-embeddings instead. When assigning meta-embeddings, most existing methods are a heuristically designed, predefined mapping from each user/item entity's ID to the corresponding meta-embedding indexes (e.g., double hashing), thus simplifying the optimization problem into learning only the meta-embeddings. However, in the context of GNN-based collaborative filtering, such a fixed mapping omits the semantic correlations between entities that are evident in the user-item interaction graph, leading to suboptimal recommendation performance. To this end, we propose Lightweight Embeddings for Graph Collaborative Filtering (LEGCF), a parameter-efficient embedding framework dedicated to GNN-based recommenders. LEGCF innovatively introduces an assignment matrix as an additional learnable component on top of meta-embeddings. To jointly optimize these two heavily entangled components, aside from learning the meta-embeddings by minimizing the recommendation loss, LEGCF further performs efficient assignment update by enforcing a novel semantic similarity constraint and finding its closed-form solution based on matrix pseudo-inverse. The meta-embeddings and assignment matrix are alternately updated, where the latter is sparsified on the fly to ensure negligible storage overhead. Extensive experiments on three benchmark datasets have verified LEGCF's smallest trade-off between size and performance, with consistent accuracy gain over state-of-the-art baselines. The codebase of LEGCF is available in https://github.com/xurong-liang/LEGCF.

Federated recommender systems (FedRecs) have gained significant attention for their potential to protect user's privacy by keeping user privacy data locally and only communicating model parameters/gradients to the server. Nevertheless, the currently existing architecture of FedRecs assumes that all users have the same 0-privacy budget, i.e., they do not upload any data to the server, thus overlooking those users who are less concerned about privacy and are willing to upload data to get a better recommendation service. To bridge this gap, this paper explores a user-governed data contribution federated recommendation architecture where users are free to take control of whether they share data and the proportion of data they share to the server. To this end, this paper presents a cloud-device collaborative graph neural network federated recommendation model, named CDCGNNFed. It trains user-centric ego graphs locally, and high-order graphs based on user-shared data in the server in a collaborative manner via contrastive learning. Furthermore, a graph mending strategy is utilized to predict missing links in the graph on the server, thus leveraging the capabilities of graph neural networks over high-order graphs. Extensive experiments were conducted on two public datasets, and the results demonstrate the effectiveness of the proposed method.

The Point-of-Interest (POI) recommendation system, designed to recommend potential future visits of users based on their check-in sequences, faces the challenge of data scarcity. This challenge primarily stems from the data sparsity issue, namely users interact with only a small number of POIs. Most existing studies attempt to solve this problem by focusing on POI check-in sequences, without considering the substantial multi-modal content information (e.g. textual and image data) commonly associated with POIs. In this paper, we propose a novel multi-modal content-aware framework for POI recommendation (MMPOI). Our approach addresses the issue of data sparsity by incorporating multi-modal content information about POIs from a new perspective. Specifically, MMPOI leverages pre-trained models for inter-modal conversion and employs a unified pre-trained model to extract modal-specific features from each modality, effectively bridging the semantic gap between different modalities. We propose to build a Multi-Modal Trajectory Flow Graph (MTFG) which combines the multi-modal semantic structure with check-in sequences. Moreover, we design an adaptive multi-task Transformer that models users' multi-modal movement patterns and integrates them for the next POI recommendation tasks. Extensive experiments on four real-world datasets demonstrate that MMPOI outperforms state-of-the-art POI recommendation methods. To facilitate reproducibility, we have released both the code and the multi-modal POI recommendation datasets we collect https://github.com/zzmylq/MMPOI

In Location-based Social Networks (LBSNs), Point-of-Interest (POI) recommendation helps users discover interesting places. There is a trend to move from the conventional cloud-based model to on-device recommendations for privacy protection and reduced server reliance. Due to the scarcity of local user-item interactions on individual devices, solely relying on local instances is not adequate. Collaborative Learning (CL) emerges to promote model sharing among users. Central to this CL paradigm is reference data, which is an intermediary that allows users to exchange their soft decisions without directly sharing their private data or parameters, ensuring privacy and benefiting from collaboration. While recent efforts have developed CL-based POI frameworks for robust and privacy-centric recommendations, they typically use a single and unified reference for all users. Reference data that proves valuable for one user might be harmful to another, given the wide range of user preferences. Some users may not offer meaningful soft decisions on items outside their interest scope. Consequently, using the same reference data for all collaborations can impede knowledge exchange and lead to sub-optimal performance. To address this gap, we introduce the Decentralized Collaborative Learning with Adaptive Reference Data (DARD) framework, which crafts adaptive reference data for effective user collaboration. It first generates a desensitized public reference data pool with transformation and probability data generation methods. For each user, the selection of adaptive reference data is executed in parallel by training loss tracking and influence function. Local models are trained with individual private data and collaboratively with the geographical and semantic neighbors. During the collaboration between two users, they exchange soft decisions based on a combined set of their adaptive reference data. Our evaluations across two real-world datasets highlight DARD's superiority in recommendation performance and addressing the scarcity of available reference data.

Traditional recommender systems estimate user preference on items purely based on historical interaction records, thus failing to capture fine-grained yet dynamic user interests and letting users receive recommendation only passively. Recent conversational recommender systems (CRSs) tackle those limitations by enabling recommender systems to interact with the user to obtain her/his current preference through a sequence of clarifying questions. Recently, there has been a rise of using knowledge graphs (KGs) for CRSs, where the core motivation is to incorporate the abundant side information carried by a KG into both the recommendation and conversation processes. However, existing KG-based CRSs are subject to two defects: (1) there is a semantic gap between the learned representations of utterances and KG entities, hindering the retrieval of relevant KG information; (2) the reasoning over KG is mostly performed with the implicitly learned user interests, overlooking the explicit signals from the entities actually mentioned in the conversation. To address these drawbacks, we propose a new CRS framework, namely, the Knowledge Enhanced Conversational Reasoning (KECR) model. As a user can reflect her/his preferences via both attribute- and item-level expressions, KECR jointly embeds the structured knowledge from two levels in the KG. A mutual information maximization constraint is further proposed for semantic alignment between the embedding spaces of utterances and KG entities. Meanwhile, KECR utilizes the connectivity within the KG to conduct explicit reasoning of the user demand, making the model less dependent on the user’s feedback to clarifying questions. As such, the semantic alignment and explicit KG reasoning can jointly facilitate accurate recommendation and quality dialogue generation. By comparing with strong baselines on two real-world datasets, we demonstrate that KECR obtains state-of-the-art recommendation effectiveness, as well as competitive dialogue generation performance.

Accurately estimating packages’ arrival time in e-commerce can enhance users’ shopping experience and improve the placement rate of products. This problem is often formalized as an Origin-Destination (OD)-based ETA (i.e., estimated time of arrival) prediction task, where the delivery time is estimated mainly based on sender and receiver addresses and other context information. One inherent challenge of the OD-based ETA problem is that the delivery time highly depends on the actual delivery trajectory which is unknown at the time of prediction. In this article, we tackle this challenge by effectively exploiting historical delivery trajectories. We propose a novel Knowledge Distillation Graph neural network-based package ETA prediction (KDG-ETA) model, which uses knowledge distillation in the training phase to distill the knowledge of historical trajectories into OD pair embeddings. In KDG-ETA, a multi-level trajectory graph representation model is proposed to fully exploit trajectory information at the node-level, edge-level, and path-level. Then, the OD representations embedded with trajectory knowledge are combined with context embeddings from feature extraction module for delivery time prediction using an adaptive attention module. KDG-ETA consistently outperforms existing state-of-the-art OD-based ETA prediction methods on three real-world Alibaba datasets, reducing the Mean Absolute Error (MAE) by 3.0%–39.1% as demonstrated in our extensive empirical evaluation.

Contrastive learning (CL) has recently been demonstrated critical in improving recommendation performance. The underlying principle of CL-based recommendation models is to ensure the consistency between representations derived from different graph augmentations of the user-item bipartite graph. This self-supervised approach allows for the extraction of general features from raw data, thereby mitigating the issue of data sparsity. Despite the effectiveness of this paradigm, the factors contributing to its performance gains have yet to be fully understood. This paper provides novel insights into the impact of CL on recommendation. Our findings indicate that CL enables the model to learn more evenly distributed user and item representations, which alleviates the prevalent popularity bias and promoting long-tail items. Our analysis also suggests that the graph augmentations, previously considered essential, are relatively unreliable and of limited significance in CL-based recommendation. Based on these findings, we put forward an eXtremelySimpleGraphContrastiveLearning method (XSimGCL) for recommendation, which discards the ineffective graph augmentations and instead employs a simple yet effective noise-based embedding augmentation to generate views for CL. A comprehensive experimental study on four large and highly sparse benchmark datasets demonstrates that, though the proposed method is extremely simple, it can smoothly adjust the uniformity of learned representations and outperforms its graph augmentation-based counterparts by a large margin in both recommendation accuracy and training efficiency. The code and used datasets are released athttps://github.com/Coder-Yu/SELFRec.

Data are invaluable in machine learning (ML), yet they raise significant privacy concerns. In the real world, data are often distributed across isolated silos, challenging conventional ML methods that centralize data. Federated learning (FL) offers a privacy-preserving solution that enables learning without direct data transfer. Meanwhile, the “right to be forgotten” sparks privacy-preserving methods from another viewpoint as machine unlearning, enabling data owners to erase specific data contributions from ML models. However, the invisibility of data in FL scenarios complicates effective local data removal, necessitating tailored unlearning algorithms for FL. Existing federated unlearning methods fall into approximate unlearning, leaving residual memorization of target data, consequently diminishing user trust. To bridge this gap, we propose FedCIO, a novel framework for exact federated unlearning, designed to efficiently manage precise data removal requests in FL scenarios. Specifically, the framework involves client clustering, isolation among clusters, and one-shot aggregation of cluster models. This framework facilitates efficient unlearning by retraining only a relevant model subset rather than from scratch. To enhance the capability to handle Non-Independent and Identically Distributed (Non-IID) data, we further introduce an advanced spectral clustering implementation based on model similarity for better cluster partitioning. Comprehensive evaluation across common FL datasets with varied distributions demonstrates the superior performance of our proposed framework.

E-commerce live streaming as an increasingly popular sales model has generated a significant amount of gross merchandise value (GMV) for e-commerce platforms. Live streaming recommendation systems (LSRS) of e-commerce aim to recommend the most appropriate live channels for users to motivate them to buy products. Existing LSRS methods focus only on the user’s interaction behaviors on the live channel (live domain) while ignoring the user’s behaviors and intentions on the e-commerce product (product domain). As a result, the user’s consistent purchase intentions in the cross-domain are not being fully captured, especially when user present differentiated purchase intentions in the cross-domain. How to disentangle user’s consistent intentions and domain-specific intentions in the cross-domain poses a challenge to the LSRS of e-commerce platforms. In this paper, we present a live channel recommendation method, named eLiveRec, developed for Taobao, one of the largest e-commerce platform in the world. Specifically, eLiveRec employs the disentangled encoder module to learn user’s cross-domain consistent intentions and domain-specific intentions. Then, an adaptive multi-task learning framework is developed to jointly optimize the multiple objectives (e.g., stay time, click goods bag, and click products after entering channel) related to live streaming recommendation. In this way, the performance of live streaming recommendation can be further improved and con-form to standard industry RS paradigms. Extensive experiments are conducted on a large-scale industry dataset collected from Taobao Live platform have been performed. Both online and offline experimental results indicate that eLiveRec consistently outperforms existing state-of-the-art baseline methods.

Predicting Estimated Time of Arrival (ETA) for packages is a critical problem in e-commerce. The prediction is often made based on spatial (sending and receiving addresses), temporal (payment time), and context (merchants) attributes. Existing methods usually formalize this task as an Origin-Destination (OD) ETA prediction problem and exploit the attribute relations with graph learning. However, most existing methods make use of fixed and manually defined graph structures, which are often not optimal for downstream ETA task and hence lead to unsatisfactory prediction results. In addition, current ETA models tend to focus on prediction accuracy without considering fulfillment rate. This may lead to a low fulfillment rate in practice, i.e., actual delivery time is much longer than estimations provided by models, which consequently exacerbates the frustrating experiences for users. To address these issues, we propose a novel Graph Structure Learning-based Quantile Regression (GSL-QR) model for e-commerce ETA prediction in this paper. Specifically, we utilize graph structure learning to dynamically update the spatial and temporal relation graphs of orders and learn optimal graph structures and graph embeddings guided by downstream ETA prediction task. To guarantee both prediction accuracy and order fulfillment rate, we design a multi-objective quantile regression in GSL-QR that can find the Pareto solution of the problem. In order to extend GSL to large-scale real-world graphs, we devise a Fast Sampling-based Graph Structure Learning (FS-GSL) method, which can significantly reduce the computational complexity of graph structure learning. Finally, we conduct comprehensive experiments on three industrial datasets collected from Alibaba e-commerce platform. The results demonstrate that the proposed model can significantly outperform baselines on both ETA prediction accuracy and order fulfillment rate.

Latent factor models are the most popular backbones for today's recommender systems owing to their prominent performance. Latent factor models represent users and items as real-valued embedding vectors for pairwise similarity computation, and all embeddings are traditionally restricted to a uniform size that is relatively large (e.g., 256-dimensional). With the exponentially expanding user base and item catalog in contemporary e commerce, this design is admittedly becoming memory-inefficient. To facilitate lightweight recommendation, reinforcement learning (RL) has recently opened up opportunities for identifying varying embedding sizes for different users/items. However, challenged by search efficiency and learning an optimal RL policy, existing RL-based methods are restricted to highly discrete, predefined embedding size choices. This leads to a largely overlooked potential of introducing finer granularity into embedding sizes to obtain better recommendation effectiveness under a given memory budget. In this paper, we propose continuous input embedding size search (CIESS), a novel RL-based method that operates on a continuous search space with arbitrary embedding sizes to choose from. In CIESS, we further present an innovative random walk-based exploration strategy to allow the RL policy to efficiently explore more candidate embedding sizes and converge to a better decision. CIESS is also model-agnostic and hence generalizable to a variety of latent factor RSs, whilst experiments on two real-world datasets have shown state-of-the-art performance of CIESS under different memory budgets when paired with three popular recommendation models.

With the accumulation of check-in data from location-based services, next Point-of-Interest (POI) recommendations are gaining increasing attention. It is well known that the spatio-temporal contextual information of user check-in behavior plays a crucial role in handling vital and inherent challenges in next POI recommendation, including capture of user dynamic preferences and the sparsity problem of check-in data. However, many studies either ignore or simply stack the context features with the embedding of POIs while relying only on POI recommendation loss to optimize the entire model, therefore failing to take full advantage of the potential information in contexts. Additionally, users’ interests are usually unstable and evolve over time, and accordingly recent studies have proposed various approaches to predict users’ next POIs by incorporating contextual information and modeling both their long- and short-term preferences, respectively. Yet many studies overemphasize the final POI recommendation performance, and the association between POI sequences and contextual information is not well embodied in data representations. In this article, we focus on the preceding problems and propose a unified attention framework for next POI recommendation by modeling users’ Long- and Short-term Preferences via Self-supervised Learning (LSPSL). Specifically, based on the self-attention network and two self-supervised optimization objectives, LSPSL first deeply exploits the intrinsic correlations between POI sequences and contextual information through pre-training, which strengthens data representations. Then, supported by pre-trained contextualized embeddings, LSPSL models and fuses users’ complex long- and short-term preferences in a unified way. Extensive experiments on real-world datasets demonstrate the superiority of our model compared with other state-of-the-art approaches.

Real-world graphs are typically distributed across multiple data centers. When performing reachability queries on these distributed graphs, reachability labeling methods ensure fast query processing by using lightweight indexes. One of the best-known labeling methods is TOL; however, TOL is a serial algorithm and cannot handle distributed graphs. The main goal of this paper is to design new labeling methods that can work in parallel while producing the same index as TOL. To this end, we investigate the limitation of TOL and thus propose a filtering-and-refinement framework for index creation. This framework first obtains a super-set of each vertex's label sets and then eliminates the invalid elements. Based on this framework, we design distributed labeling algorithms and then use batch processing to improve efficiency. Experimental results on real-world graphs show that the proposed algorithms can index distributed graphs efficiently.

Contrastive learning (CL) recently has spurred a fruitful line of research in the field of recommendation, since its ability to extract self-supervised signals from the raw data is well-aligned with recommender systems' needs for tackling the data sparsity issue. A typical pipeline of CL-based recommendation models is first augmenting the user-item bipartite graph with structure perturbations, and then maximizing the node representation consistency between different graph augmentations. Although this paradigm turns out to be effective, what underlies the performance gains is still a mystery. In this paper, we first experimentally disclose that, in CL-based recommendation models, CL operates by learning more uniform user/item representations that can implicitly mitigate the popularity bias. Meanwhile, we reveal that the graph augmentations, which used to be considered necessary, just play a trivial role. Based on this finding, we propose a simple CL method which discards the graph augmentations and instead adds uniform noises to the embedding space for creating contrastive views. A comprehensive experimental study on three benchmark datasets demonstrates that, though it appears strikingly simple, the proposed method can smoothly adjust the uniformity of learned representations and has distinct advantages over its graph augmentation-based counterparts in terms of recommendation accuracy and training efficiency. The code is released at https://github.com/Coder-Yu/QRec.

Being a fundamental graph operator, reachability query has been widely studied by the data mining community in the past decades. In a directed acyclic graph (DAG), one vertex is reachable by another if there exists a chain of directed edges connecting the two vertexes. The state-of-the-art (SOTA) reachability query methods mostly first index all the vertexes in the underlying DAG and assign them with different labels, and then use these indexes and/or labels to efficiently filter out as many unreachable queries as possible. Thus, because a large portion of unreachable queries can be identified without evoking any tedious path-finding process, the overall time taken by a huge number of queries is much shortened with a tolerable compensation on the additional index and/or label preprocessing time and space. In this paper, we propose the Extreme Labeling Filter (ELF), which is a novel generic filter that can be applied to existing reachability query methods to additionally identify a large number of unreachable queries. Based on the analysis of the given DAG in a systematic and autonomous manner, ELF first determines whether to use predecessors or successors to label the vertexes. Based on such self-determined labels, ELF is then able to identify a large number of unreachable queries with a low time complexity of O(1). To evaluate the performance of ELF, we apply it on 4 reachability query methods (1 conventional and 3 SOTA, all designated for reachability query in DAGs) and conduct experiments on 17 datasets of different sizes. The experimental results show that by applying ELF, all methods significantly shorten the query time.

Due to the growing privacy concerns, decentralization emerges rapidly in personalized services, especially recommendation. Also, recent studies have shown that centralized models are vulnerable to poisoning attacks, compromising their integrity. In the context of recommender systems, a typical goal of such poisoning attacks is to promote the adversary's target items by interfering with the training dataset and/or process. Hence, a common practice is to subsume recommender systems under the decentralized federated learning paradigm, which enables all user devices to collaboratively learn a global recommender while retaining all the sensitive data locally. Without exposing the full knowledge of the recommender and entire dataset to end-users, such federated recommendation is widely regarded 'safe' towards poisoning attacks. In this paper, we present a systematic approach to backdooring federated recommender systems for targeted item promotion. The core tactic is to take advantage of the inherent popularity bias that commonly exists in data-driven recommenders. As popular items are more likely to appear in the recommendation list, our innovatively designed attack model enables the target item to have the characteristics of popular items in the embedding space. Then, by uploading carefully crafted gradients via a small number of malicious users during the model update, we can effectively increase the exposure rate of a target (unpopular) item in the resulted federated recommender. Evaluations on two real-world datasets show that 1) our attack model significantly boosts the exposure rate of the target item in a stealthy way, without harming the accuracy of the poisoned recommender; and 2) existing defenses are not effective enough, highlighting the need for new defenses against our local model poisoning attacks to federated recommender systems.

Question Generation (QG), as a challenging Natural Language Processing task, aims at generating questions based on given answers and context. Existing QG methods mainly focus on building or training models for specific QG datasets. These works are subject to two major limitations: (1) They are dedicated to specific QG formats (e.g., answer-extraction or multi-choice QG), therefore, if we want to address a new format of QG, a re-design of the QG model is required. (2) Optimal performance is only achieved on the dataset they were just trained on. As a result, we have to train and keep various QG models for different QG datasets, which is resource-intensive and ungeneralizable.

Abstract Nowadays recording and sharing personal lives using mobile devices on the Internet is becoming increasingly popular, and successive POI recommendation is gaining growing attention from academia and industry. In mobile scenarios, multiple influencing factors including the diversity of user preferences, the changeability of user behavior and the dynamic of spatiotemporal context bring great challenges to the POI recommender system. In order to accurately capture both the stable and the contextual preferences of mobile users in dynamic contexts, we propose a fusion framework JANICP (Joint Attention Networks with Inherent and Contextual Preferences) for successive POI recommendation by jointly training an offline/nearline user inherent interest perception model and an online user contextual interest prediction model. The offline model is trained based on the global historical behavior data to achieve stable interest representation, while the online model is trained based on the instantly selected context-sensitive data to achieve dynamic interest perception. An attention aggregation and matching module is used to fully connect the two kinds of preference representations and generate the final POI recommendation. Extensive experiments were conducted on three real datasets and experimental results show that the proposed JANICP outperforms existing state-of-the-art methods.

We present a novel approach for estimating the proportional distributions of function types (i.e. functional distributions) in an urban area through learning semantics preserved embeddings of points-of-interest (POIs). Specifically, we represent POIs as low-dimensional vectors to capture (1) the spatial co-occurrence patterns of POIs and (2) the semantics conveyed by the POI hierarchical categories (i.e. categorical semantics). The proposed approach utilizes spatially explicit random walks in a POI network to learn spatial co-occurrence patterns, and a manifold learning algorithm to capture categorical semantics. The learned POI vector embeddings are then aggregated to generate regional embeddings with long short-term memory (LSTM) and attention mechanisms, to take account of the different levels of importance among the POIs in a region. Finally, a multilayer perceptron (MLP) maps regional embeddings to functional distributions. A case study in Xiamen Island, China implements and evaluates the proposed approach. The results indicate that our approach outperforms several competitive baseline models in all evaluation measures, and yields a relatively high consistency between the estimation and ground truth. In addition, a comprehensive error analysis unveils several intrinsic limitations of POI data for this task, e.g. ambiguous linkage between POIs and functions.

Federated Learning (FL) bridges the gap between collaborative machine learning and preserving data privacy. To sustain the long-term operation of an FL ecosystem, it is important to attract high-quality data owners with appropriate incentive schemes. As an important building block of such incentive schemes, it is essential to fairly evaluate participants’ contribution to the performance of the final FL model without exposing their private data. Shapley Value (SV)–based techniques have been widely adopted to provide a fair evaluation of FL participant contributions. However, existing approaches incur significant computation costs, making them difficult to apply in practice. In this article, we propose the Guided Truncation Gradient Shapley (GTG-Shapley) approach to address this challenge. It reconstructs FL models from gradient updates for SV calculation instead of repeatedly training with different combinations of FL participants. In addition, we design a guided Monte Carlo sampling approach combined with within-round and between-round truncation to further reduce the number of model reconstructions and evaluations required. This is accomplished through extensive experiments under diverse realistic data distribution settings. The results demonstrate that GTG-Shapley can closely approximate actual Shapley values while significantly increasing computational efficiency compared with the state-of-the-art, especially under non-i.i.d. settings.

Social recommender systems are expected to improve recommendation quality by incorporating social information when there is little user-item interaction data. However, recent reports from industry show that social recommender systems consistently fail in practice. According to the negative findings, the failure is attributed to: (1) A majority of users only have a very limited number of neighbors in social networks and can hardly benefit from social relations; (2) Social relations are noisy but they are indiscriminately used; (3) Social relations are assumed to be universally applicable to multiple scenarios while they are actually multi-faceted and show heterogeneous strengths in different scenarios. Most existing social recommendation models only consider the homophily in social networks and neglect these drawbacks. In this paper we propose a deep adversarial framework based on graph convolutional networks (GCN) to address these problems. Concretely, for (1) and (2), a GCN-based autoencoder is developed to augment the relation data by encoding high-order and complex connectivity patterns, and meanwhile is optimized subject to the constraint of reconstructing the social profile to guarantee the validity of the identified neighborhood. After obtaining enough purified social relations for each user, a GCN-based attentive social recommendation module is designed to address (3) by capturing the heterogeneous strengths of social relations. Finally, we adopt adversarial training to unify all the components by playing a Minimax game and ensure a coordinated effort to enhance recommendation performance. Extensive experiments on multiple open datasets demonstrate the superiority of our framework and the ablation study confirms the importance and effectiveness of each component.

Hypergraphs have been becoming a popular choice to model complex, non-pairwise, and higher-order interactions for recommender systems. However, compared with traditional graph-based methods, the constructed hypergraphs are usually much sparser, which leads to a dilemma when balancing the benefits of hypergraphs and the modelling difficulty. Moreover, existing sequential hypergraph recommendation overlooks the temporal modelling among user relationships, which neglects rich social signals from the recommendation data. To tackle the above shortcomings of the existing hypergraph-based sequential recommendations, we propose a novel architecture named Hyperbolic Hypergraph representation learning method for Sequential Recommendation (H2SeqRec) with the pre-training phase. Specifically, we design three self-supervised tasks to obtain the pre-training item embeddings to feed or fuse into the following recommendation architecture (with two ways to use the pre-trained embeddings). In the recommendation phase, we learn multi-scale item embeddings via a hierarchical structure to capture multiple time-span information. To alleviate the negative impact of sparse hypergraphs, we utilize a hyperbolic space-based hypergraph convolutional neural network to learn the dynamic item embeddings. Also, we design an item enhancement module to capture dynamic social information at each timestamp to improve effectiveness. Extensive experiments are conducted on two real-world datasets to prove the effectiveness and high performance of the model.

Session-based recommendation targets next-item prediction by exploiting user behaviors within a short time period. Compared with other recommendation paradigms, session-based recommendation suffers more from the problem of data sparsity due to the very limited short-term interactions. Self-supervised learning, which can discover ground-truth samples from the raw data, holds vast potentials to tackle this problem. However, existing self-supervised recommendation models mainly rely on item/segment dropout to augment data, which are not fit for session-based recommendation because the dropout leads to sparser data, creating unserviceable self-supervision signals. In this paper, for informative session-based data augmentation, we combine self-supervised learning with co-training, and then develop a framework to enhance session-based recommendation. Technically, we first exploit the session-based graph to augment two views that exhibit the internal and external connectivities of sessions, and then we build two distinct graph encoders over the two views, which recursively leverage the different connectivity information to generate ground-truth samples to supervise each other by contrastive learning. In contrast to the dropout strategy, the proposed self-supervised graph co-training preserves the complete session information and fulfills genuine data augmentation. Extensive experiments on multiple benchmark datasets show that, session-based recommendation can be remarkably enhanced under the regime of self-supervised graph co-training, achieving the state-of-the-art performance.

Crowdsourcing is a popular and relatively economic way to harness human intelligence to process computer-hard tasks. Due to diverse factors (i.e., task difficulty, worker capability, and incentives), the collected answers from various crowd workers are of different quality. Many approaches have been proposed to manage high quality answers and to reduce the budget by modelling tasks, workers, or both. However, most of the existing approaches implicitly assume that the capability of workers is fixed during the crowdsourcing process. But in practice, such capability can be improved by gradually completing easy to hard tasks, alike human beings’ intrinsic self-paced learning ability. In this paper, we investigate crowdsourcing with self-paced workers, whose capability can be gradually boosted as he/she scrutinises and completes easy to hard tasks. Our proposed SPCrowd (Self-Paced Crowd worker) first asks workers to complete a set of golden tasks with known annotations; provides feedback to assist workers with capturing the raw modes of tasks and to spark the self-paced learning, which in turn facilitates the estimation of workers’ quality and tasks’ difficulty. It then introduces a task difficulty model to quantify the difficulty of tasks and rank them from easy to hard, and a benefit maximization criterion for task assignment, which can dynamically monitor the quality of self-paced workers and assign the sorted tasks to capable workers. In this way, a worker can successfully complete hard tasks after he/she completes easier and related tasks. Experimental results on semi-simulated and real crowdsourcing projects show that SPCrowd can better control the quality and save the budget compared to competitive baselines.

Partial multi-label learning (PML) aims at learning a robust multi-label classifier by training on ambiguous data, where each sample is associated with a set of candidate labels, among which only a subset are valid labels. A basic premise of existing PML solutions is to obtain enough partial multi-label samples for inducing the classification model. However, when dealing with new tasks, we may only have a few PML samples for those tasks. Furthermore, existing few-shot learning approaches assume the support (training) samples are precisely labeled; as such, irrelevant labels in the candidate label set may seriously mislead the meta-learner and thus result in a compromised performance. How to achieve PML with limited few-shot support samples is an important and practical problem, but not yet well studied. In this paper, we propose an approach called FsPML (Few-shot PML) to tackle this problem. Specifically, FsPML first performs adaptive distance metric learning via an embedding network using both sample features and label semantics in the embedding space. Next it rectifies the positive and negative prototypes of each new label of the target task in the embedding space. An unseen example can then be classified via its distances to the positive and to the negative prototypes. Experimental results on widely-used multi-label datasets (MS COCO and NUS-WIDE) demonstrate that our FsPML outperforms competitive baselines across different settings, and it can quickly generalize to new tasks with fewer training samples.

Hyperbolic space and hyperbolic embeddings are becoming a popular research field for recommender systems. However, it is not clear under what circumstances the hyperbolic space should be considered. To fill this gap, This paper provides theoretical analysis and empirical results on when and where to use hyperbolic space and hyperbolic embeddings in recommender systems. Specifically, we answer the questions that which type of models and datasets are more suited for hyperbolic space, as well as which latent size to choose. We evaluate our answers by comparing the performance of Euclidean space and hyperbolic space on different latent space models in both general item recommendation domain and social recommendation domain, with 6 widely used datasets and different latent sizes. Additionally, we propose a new metric learning based recommendation method called SCML and its hyperbolic version HSCML. We evaluate our conclusions regarding hyperbolic space on SCML and show the state-of-the-art performance of hyperbolic space by comparing HSCML with other baseline methods.

In recent years, recommender systems play a pivotal role in helping users identify the most suitable items that satisfy personal preferences. As user-item interactions can be naturally modelled as graph-structured data, variants of graph convolutional networks (GCNs) have become a well-established building block in the latest recommenders. Due to the wide utilization of sensitive user profile data, existing recommendation paradigms are likely to expose users to the threat of privacy breach, and GCN-based recommenders are no exception. Apart from the leakage of raw user data, the fragility of current recommenders under inference attacks offers malicious attackers a backdoor to estimate users’ private attributes via their behavioral footprints and the recommendation results. However, little attention has been paid to developing recommender systems that can defend such attribute inference attacks, and existing works achieve attack resistance by either sacrificing considerable recommendation accuracy or only covering specific attack models or protected information. In our paper, we propose GERAI, a novel differentially private graph convolutional network to address such limitations. Specifically, in GERAI, we bind the information perturbation mechanism in differential privacy with the recommendation capability of graph convolutional networks. Furthermore, based on local differential privacy and functional mechanism, we innovatively devise a dual-stage encryption paradigm to simultaneously enforce privacy guarantee on users’ sensitive features and the model optimization process. Extensive experiments show the superiority of GERAI in terms of its resistance to attribute inference attacks and recommendation effectiveness.

Abstract Microtask crowdsourcing is a form of crowdsourcing in which work is decomposed into a set of small, self-contained tasks, which each can typically be completed in a matter of minutes. Due to the various capabilities and knowledge background of the voluntary participants on the Internet, the answers collected from the crowd are ambiguous and the final answer aggregation is challenging. In this process, the choice of quality control strategies is important for ensuring the quality of the crowdsourcing results. Previous work on answer estimation mainly used expectation–maximization (EM) approach. Unfortunately, EM provides local optimal solutions and the estimated results will be affected by the initial value. In this paper, we extend the local optimal result of EM and propose an approximate global optimal algorithm for answer aggregation of crowdsourcing microtasks with binary answers. Our algorithm is expected to improve the accuracy of real answer estimation through further likelihood maximization. First, three worker quality evaluation models are presented based on static and dynamic methods, respectively, and the local optimal results are obtained based on the maximum likelihood estimation method. Then, a dominance ordering model (DOM) is proposed according to the known worker responses and worker categories for the specified crowdsourcing task to reduce the space of potential task-response sequence while retaining the dominant sequence. Subsequently, a Cut-point neighbor detection algorithm is designed to iteratively search for the approximate global optimal estimation in a reduced space, which works on the proposed dominance ordering model (DOM). We conduct extensive experiments on both simulated and real-world datasets, and the experimental results illustrate that the proposed approach can obtain better estimation results and has higher performance than regular EM-based algorithms.

Emotion recognition has become an important component of human–computer interaction systems. Research on emotion recognition based on electroencephalogram (EEG) signals are mostly conducted by the analysis of all channels' EEG signals. Although some progresses are achieved, there are still several challenges such as high dimensions, correlation between different features and feature redundancy in the realistic experimental process. These challenges have hindered the applications of emotion recognition to portable human–computer interaction systems (or devices). This paper explores how to find out the most effective EEG features and channels for emotion recognition so as to only collect data as less as possible. First, discriminative features of EEG signals from different dimensionalities are extracted for emotion classification, including the first difference, multiscale permutation entropy, Higuchi fractal dimension, and discrete wavelet transform. Second, relief algorithm and floating generalized sequential backward selection algorithm are integrated as a novel channel selection method. Then, support vector machine is employed to classify the emotions for verifying the performance of the channel selection method and extracted features. At last, experimental results demonstrate that the optimal channel set, which are mostly located at the frontal, has extremely high similarity on the self-collected data set and the public data set and the average classification accuracy is achieved up to 91.31% with the selected 10-channel EEG signals. The findings are valuable for the practical EEG-based emotion recognition systems.

Group Recommendation (GR) is the task of suggesting relevant items/events for a group of users in online systems, whose major challenge is to aggregate the preferences of group members to infer the decision of a group. Prior group recommendation methods applied predefined static strategies for preference aggregation. However, these static strategies are insufficient to model the complicated decision making process of a group, especially for occasional groups which are formed adhoc. Compared to conventional individual recommendation task, GR is rather dynamic and each group member may contribute differently to the final group decision. Recent works argue that group members should have non-uniform weights in forming the decision of a group, and try to utilize a standard attention mechanism to aggregate the preferences of group members, but they do not model the interaction behavior among group members, and the decision making process is largely unexplored.In this work, we study GR in a more general scenario, that is Occasional Group Recommendation (OGR), and focus on solving the preference aggregation problem and the data sparsity issue of group-item interactions. Instead of exploring new heuristic or vanilla attention-based mechanism, we propose a new social self-attention based aggregation strategy by directly modeling the interactions among group members, namely Group Self-Attention (GroupSA). In GroupSA, we treat the group decision making process as multiple voting processes, and develop a stacked social self-attention network to simulate how a group consensus is reached. To overcome the data sparsity issue, we resort to the relatively abundant user-item and user-user interaction data, and enhance the representation of users by two types of aggregation methods. In the training process, we further propose a joint training method to learn the user/item embeddings in the group-item recommendation task and the user-item recommendation task simultaneously. Finally, we conduct extensive experiments on two real-world datasets. The experimental results demonstrate the superiority of our proposed GroupSA method compared to several state-of-the-art methods in terms of HR and NDCG.

In recent years, recommender system has become an indispensable function in all e-commerce platforms. The review rating data for a recommender system typically comes from open platforms, which may attract a group of malicious users to deliberately insert fake feedback in an attempt to bias the recommender system to their favour. The presence of such attacks may violate modeling assumptions that high-quality data is always available and these data truly reflect users' interests and preferences. Therefore, it is of great practical significance to construct a robust recommender system that is able to generate stable recommendations even in the presence of shilling attacks. In this paper, we propose GraphRfi - a GCN-based user representation learning framework to perform robust recommendation and fraudster detection in a unified way. In its end-to-end learning process, the probability of a user being identified as a fraudster in the fraudster detection component automatically determines the contribution of this user's rating data in the recommendation component; while the prediction error outputted in the recommendation component acts as an important feature in the fraudster detection component. Thus, these two components can mutually enhance each other. Extensive experiments have been conducted and the experimental results show the superiority of our GraphRfi in the two tasks - robust rating prediction and fraudster detection. Furthermore, the proposed GraphRfi is validated to be more robust to the various types of shilling attacks over the state-of-the-art recommender systems.

As considerable amounts of POI check-in data have been accumulated, successive point-of-interest (POI) recommendation is increasingly popular. Existing successive POI recommendation methods only predict where user will go next, ignoring when this behavior will occur. In this work, we focus on predicting POIs that will be visited by users in the next 24 hours. As check-in data is very sparse, it is challenging to accurately capture user preferences in temporal patterns. To this end, we propose a category-aware deep model CatDM that incorporates POI category and geographical influence to reduce search space to overcome data sparsity. We design two deep encoders based on LSTM to model the time series data. The first encoder captures user preferences in POI categories, whereas the second exploits user preferences in POIs. Considering clock influence in the second encoder, we divide each user’s check-in history into several different time windows and develop a personalized attention mechanism for each window to facilitate CatDM to exploit temporal patterns. Moreover, to sort the candidate set, we consider four specific dependencies: user-POI, user-category, POI-time and POI-user current preferences. Extensive experiments are conducted on two large real datasets. The experimental results demonstrate that our CatDM outperforms the state-of-the-art models for successive POI recommendation on sparse check-in data.

The information of the real world is stored as triplets (head entity, relation, tail entity) in knowledge graphs. They are extremely useful resources for many intelligent applications but suffer from incompleteness. This paper proposes a knowledge graph representation model to infer latent privacy based on the existing data in attribute network. In our model, considering the nodes are heterogeneous, we classify the nodes into attribute nodes and entity nodes. In order to protect the privacy of entities, we don't follow the previous methods to learn and store the feature embedding of each entity in knowledge graph. Our model focuses in capturing the restriction patterns of attribute nodes, which is safe when merging data from various sources. Given a triplet (entity node, relation, attribute node), firstly, we get the embedding of the entity node by using a sophisticated way to utilize all the information of the node, not only the node connections but also the external text information. Then, we infer the attribute node for the entity node in a certain relation. Finally, we calculate the probability that the triplet is exist. In experiments, we evaluate our model on the tasks of triplet classification and link prediction. Evaluation results show that our approach outperforms the state-of-the-art methods with an accuracy rate of 90.0% in the task of triplet classification on the person attribute knowledge graph FB13. Besides, our model reaches promising performance by MeanRank =5.10, Hits@l = 35.14% and Hits@5=64.94% in the task of conference prediction on the academic network DBLP.

Video contents have become a critical tool for promoting products in E-commerce. However, the lack of automatic promotional video generation solutions makes large-scale video-based promotion campaigns infeasible. The first step of automatically producing promotional videos is to generate visual storylines, which is to select the building block footage and place them in an appropriate order. This task is related to the subjective viewing experience. It is hitherto performed by human experts and thus, hard to scale. To address this problem, we propose WundtBackpack, an algorithmic approach to generate storylines based on available visual materials, which can be video clips or images. It consists of two main parts, 1) the Learnable Wundt Curve to evaluate the perceived persuasiveness based on the stimulus intensity of a sequence of visual materials, which only requires a small volume of data to train; and 2) a clustering-based backpacking algorithm to generate persuasive sequences of visual materials while considering video length constraints. In this way, the proposed approach provides a dynamic structure to empower artificial intelligence (AI) to organize video footage in order to construct a sequence of visual stimuli with persuasive power. Extensive real-world experiments show that our approach achieves close to 10% higher perceived persuasiveness scores by human testers, and 12.5% higher expected revenue compared to the best performing state-of-the-art approach.

Collaborative filtering (CF) has become one of the most popular and widely used methods in recommender systems, but its performance degrades sharply for users with rare interaction data. Most existing hybrid CF methods try to incorporate side information such as review texts to alleviate the data sparsity problem. However, the process of exploiting and integrating side information is computationally expensive. Existing hybrid recommendation methods treat each user equally and ignore that the pure CF methods have already achieved both effective and efficient recommendation performance for active users with sufficient interaction records and the little improvement brought by side information to these active users is ignorable. Therefore, they are not cost-effective solutions. One cost-effective idea to bypass this dilemma is to generate sufficient "real" interaction data for the inactive users with the help of side information, and then a pure CF method could be performed on this augmented dataset effectively. However, there are three major challenges to implement this idea. Firstly, how to ensure the correctness of the generated interaction data. Secondly, how to combine the data augmentation process and recommendation process into a unified model and train the model end-to-end. Thirdly, how to make the solution generalizable for various side information and recommendation tasks. In light of these challenges, we propose a generic and effective CF model called AugCF that supports a wide variety of recommendation tasks. AugCF is based on Conditional Generative Adversarial Nets that additionally consider the class (like or dislike) as a feature to generate new interaction data, which can be a sufficiently real augmentation to the original dataset. Also, AugCF adopts a novel discriminator loss and Gumbel-Softmax approximation to enable end-to-end training. Finally, extensive experiments are conducted on two large-scale recommendation datasets, and the experimental results show the superiority of our proposed model.

The users of Twitter-like social media normally use the "@'' sign to select a suitable person to mention. It is a significant role in promoting the user experience and information propagation. To help users easily find the usernames they want to mention, the mention recommendation task has received considerable attention in recent years. Previous methods only incorporated textual information when performing this task. However, many users not only post texts on social media but also the corresponding images. These images can provide additional information that is not included in the text, which could be helpful in improving the accuracy of a mention recommendation. To make full use of textual and visual information, we propose a novel cross-attention memory network to perform the mention recommendation task for multimodal tweets. We incorporate the interests of users with external memory and use the cross-attention mechanism to extract both textual and visual information. Experimental results on a dataset collected from Twitter demonstrated that the proposed method can achieve better performance than state-of-the-art methods that use textual information only.

Computing $$k\text {-}core$$ s is a fundamental and important graph problem, which can be applied in many areas, such as community detection, network visualization, and network topology analysis. Due to the complex relationship between different entities, dual graph widely exists in the applications. A dual graph contains a physical graph and a conceptual graph, both of which have the same vertex set. Given that there exist no previous studies on the $$k\text {-}core$$ in dual graphs, we formulate a k-connected core ( $$k\text {-}CCO$$ ) model in dual graphs. A $$k\text {-}CCO$$ is a $$k\text {-}core$$ in the conceptual graph, and also connected in the physical graph. Given a dual graph and an integer k, we propose a polynomial time algorithm for computing all $$k\text {-}CCO$$ s. We also propose three algorithms for computing all maximum-connected cores ( $$MCCO$$ ), which are the existing $$k\text {-}CCO$$ s such that a $$(k+1)$$ - $$CCO$$ does not exist. We further study a subgraph search problem, which is computing a $$k\text {-}CCO$$ that contains a set of query vertices. We propose an index-based approach to efficiently answer the query for any given parameter k. We conduct extensive experiments on six real-world datasets and four synthetic datasets. The experimental results demonstrate the effectiveness and efficiency of our proposed algorithms.

Today, Pareto-optimal objects finding has been applied in various fields, such as group decision making and opinion collection. Many of the existing solutions to this problem require explicit attributes for objects. However, these attributes cannot be obtained sometimes. To address this issue, we propose an algorithm, which uses preference relations given by crowdsourcing, to find Pareto-optimal objects with shorter latency and lower monetary costs. It employs two multi-pairwise-comparison question models: BEST-form and BETTER-form questions. Multiple BEST (or BETTER) questions can be sent to crowds concurrently. Extensive experimental results show that the number of questions reduces greatly. In addition, the numerical results show that the latency is significantly shortened at a reasonable monetary cost, compared with the existing methods.

The rapid growth of data makes it possible for us to study human behavior patterns. Knowing the patterns of human behavior is of great use to help us detect the unusual fraud human behavior. Existing fraud detection methods can be divided into two categories: pattern based and outlier detection based methods. However, because of the sparsity and complex granularity of big data, these methods have high false positive in fraud detection. In this paper, we propose an effective hybrid fraud detection method. We propose SSIsomap which improves isomap to cluster behaviors into behavior classes and propose SimLOF which improves LOF to conduct outlier detection, then we use Dempster-Shafer evidence Theory for combining behavior pattern evidence and outlier evidence, which yields a degree of belief of fraud to the new coming claim. The experiment result shows our method has significantly higher accuracy than exsiting methods in medical insurance fraud detection.