Recent years have witnessed the prosperity of Knowledge Graph-Based Recommendation System (KGRS), which enriches the representation of users, items, and entities by structural knowledge with striking improvement. Nevertheless, its unaffordable computational cost still limits researchers from exploring more sophisticated models. We observe that the bottleneck for training efficiency arises from the knowledge graph, which is plagued by the well-known issue of knowledge explosion. Recently, some works have attempted to slim the inflated KG via summarization techniques, which summarize multiple real nodes into the single virtual one. However, these summarized virtual nodes may ignore collaborative signals and thus fail to figure out the redundant nodes related to recommendation task. To this end, in this article, we propose a novel approach called KGTrimmer for knowledge graph pruning tailored for recommendation, to remove the unessential nodes while minimizing performance degradation. Specifically, we design an importance evaluator from a dual-view perspective. For the collective view, we embrace the idea of collective intelligence by extracting community consensus based on abundant collaborative signals, i.e., nodes are considered important if they attract attention of numerous users. For the holistic view, we learn a global mask to identify the valueless nodes from their inherent properties or overall popularity. With the collective and holistic importance scores, we build an end-to-end importance-aware graph neural network, which injects filtered knowledge to enhance the distillation of valuable user-item collaborative signals. Ultimately, we generate a pruned knowledge graph with lightweight, stable, and robust properties to facilitate the following-up recommendation task. Extensive experiments are conducted on three publicly available datasets to prove the effectiveness and generalizability of KGTrimmer, where it can reduce the number of triplets in KG by up to 90% without compromising performance.

Post-Click Conversion Rate (CVR) prediction aims to predict the probability of a conversion event occurring after a user clicks. Most CVR prediction methods use clicked events to train models and subsequently predict on both clicked and unclicked events, facing selection bias. To unbiasedly predict CVR, doubly robust (DR) learning incorporates propensity score reweighting and missing data error imputation, but with suboptimal performance under sparse click events. We theoretically demonstrate that existing DR methods face high or even unbounded bias, variance, and generalization error bound under small propensity scores from sparse click events. This motivates us to propose a new Batch-Adaptive DR (BADR) Learning method. In particular, we propose a BADR estimator, which adaptively adjusts the influence of each data batch during debiasing CVR prediction model training based on the propensity scores within that batch. We prove that the BADR estimator has bounded bias, variance, and generalization error bound, all of which are smaller than those of the DR estimator under small propensity scores, while maintaining asymptotic double robustness, i.e., achieving double robustness under a large sample size. Furthermore, we approximate the variance of the BADR estimator and derive a new batch-adaptive imputation model training loss compatible with the BADR estimator, which theoretically ensures further variance reduction during training. Our experiments on real-world datasets validate BADR’s effectiveness and rationality.

Recommendation systems have become ubiquitous tools in online platforms, providing personalized suggestions based on user–item interactions. To capture the dynamic higher-order connections between users and items, recommendation approaches based on dynamic graphs have garnered significant attention from researchers. However, existing recommendation methods based on dynamic graphs are often limited by data sparsity, which prevents them from achieving satisfactory performance. Fortunately, the rapid development of large language models (LLMs) with powerful text generation capabilities and extensive domain knowledge has offered new possibilities for addressing this challenge. However, how to effectively integrate LLMs with dynamic graphs remains unexplored. To bridge this gap, in this article, we propose a novel framework, that is, DynLLM, for applying LLMs to dynamic graph-based recommendation methods. Specifically, DynLLM harnesses the power of LLMs to generate multi-faceted user profiles based on the rich textual features of historical purchase records, which in turn supplement and enrich the underlying relationships between users and items. Along this line, to fuse the multi-faceted profiles with temporal graph embedding, we engage LLMs to derive corresponding profile embeddings and further employ a distilled attention mechanism to refine the LLM-generated profile embeddings for alleviating noisy signals, while also assessing and adjusting the relevance of each distilled facet embedding for seamless integration with temporal graph embedding from continuous time dynamic graphs (CTDGs). Extensive experiments on three real datasets have validated the superior improvements of DynLLM over a wide range of state-of-the-art baseline methods. The implementation code is available online at https://github.com/meteor-gif/DynLLM .

A well-developed recommendation system can not only leverage multi-typed interactions (such as page view , add-to-cart , and purchase ) to better identify user preferences but also demonstrate high performance, low complexity, and strong interpretability. However, many existing solutions for multi-behavior recommendation fall short of intuitive modeling of real-world scenarios, leading to overly complex models with massive parameters and cumbersome components. In particular, they share two critical limitations: (1) Some pioneering models are built upon the strict assumption of cascade effects across behaviors, which contradicts multifarious behavior paths in practical applications. (2) Existing approaches fail to explicitly capture the unique idiosyncrasies of users and even neglect the inherent nature of items involved in the multi-behavior interactions. To this end, we propose a novel Directed Acyclic Graph Convolutional Network (DA-GCN) for the multi-behavior recommendation task. Specifically, we pinpoint the partial order relations within the monotonic behavior chain and extend it to personalized directed acyclic behavior graphs to exploit behavior dependencies. Then, a GCN-based directed edge encoder is employed to distill rich collaborative signals embodied by each directed edge. In light of the information flows over the directed acyclic structure, we propose an attentive aggregation module to gather messages from all potential antecedent behaviors, representing distinct perspectives to understand the terminated behavior. Thus, we obtain comprehensive representations for the follow-up behavior through learnable distributions over its preceding behaviors, explicitly reflecting personalized interactive patterns of users and underlying properties of items simultaneously. Finally, we design a customized multi-task learning objective for flexible joint optimization. Extensive experiments on public benchmarking datasets fully demonstrate the superiority of DA-GCN with significant performance improvement and computational efficiency over a wide range of state-of-the-art methods. Our code is available at https://github.com/xizhu1022/DA-GCN .

Last year has witnessed the re-flourishment of tag-aware recommender systems supported by the LLM-enriched tags. Unfortunately, though large efforts have been made, current solutions may fail to describe the diversity and uncertainty inherent in user preferences with only tag-driven profiles. Recently, with the development of geometry-based techniques, e.g., box embeddings, the diversity of user preferences now could be fully modeled as the range within a box in high dimension space. However, defect still exists as these approaches are incapable of capturing high-order neighbor signals, i.e., semantic-rich multi-hop relations within the user-tag-item tripartite graph, which severely limits the effectiveness of user modeling. To deal with this challenge, in this paper, we propose a novel framework, called BoxGNN, to perform message aggregation via combinations of logical operations, thereby incorporating high-order signals. Specifically, we first embed users, items, and tags as hyper-boxes rather than simple points in the representation space, and define two logical operations, i.e., union and intersection, to facilitate the subsequent process. Next, we perform the message aggregation mechanism via the combination of logical operations, to obtain the corresponding high-order box representations. Finally, we adopt a volume-based learning objective with Gumbel smoothing techniques to refine the representation of boxes. Extensive experiments on two publicly available datasets and one LLM-enhanced e-commerce dataset have validated the superiority of BoxGNN compared with various state-of-the-art baselines. The code is released online: https://github.com/critical88/BoxGNN.

Graph learning methods have achieved remarkable performance in various domains such as social recommendation, financial fraud detection, and so on. In real applications, the underlying graph is often dynamically evolving and thus, some recent studies focus on integrating the temporal topology information of graphs into the GNN for learning graph embedding. However, the robustness of training GNNs for dynamic graphs has not been discussed so far. The major reason is how to attack dynamic graph embedding still remains largely untouched, let alone how to defend against the attacks. To enable robust training of GNNs for dynamic graphs, in this paper, we investigate the problem of how to generate attacks and defend against attacks for dynamic graph embedding. Attacking dynamic graph embedding is more challenging than attacking static graph embedding as we need to understand the temporal dynamics of graphs as well as its impact on the embedding and the injected perturbations should be distinguished from the natural evolution. In addition, the defense is very challenging as the perturbations may be hidden within the natural evolution. To tackle these technical challenges, in this paper, we first develop a novel gradient-based attack method from an optimization perspective to generate perturbations to fool dynamic graph learning methods, where a key idea is to use gradient dynamics to attack the natural dynamics of the graph. Further, we borrow the idea of the attack method and integrate it with adversarial training to train a more robust dynamic graph learning method to defend against hand-crafted attacks. Finally, extensive experiments on two real-world datasets demonstrate the effectiveness of the proposed attack and defense method, where our defense method not only achieves comparable performance on clean graphs but also significantly increases the defense performance on attacked graphs.

Dynamic share recommendation, which aims at recommending a friend who would like to share a particular item at a certain timestamp, has emerged as a novel task for social-oriented e-commerce platforms. Different from traditional graph-based recommendation tasks, with integrating the interconnected social interactions and fine-grained temporal information from historical share records, this novel task may encounter one unique challenge, i.e., how to deal with the dynamic social connections and asymmetric share interactions. Even worse, users may keep inactive during some periods, which results in difficulties in updating personalized profiles. To address the above challenges, in this paper, we propose a dynamic graph share recommendation model called DynShare. Specifically, we first divide each user embedding into two parts, namely the invitation embedding and vote embedding to show the tendencies of sending and receiving items, respectively. Then, temporal graph attention networks (TGATs) based on bi-directional continuous time dynamic graphs (CTDGs) are leveraged to encode temporal neighbor information from different directions. Afterward, to estimate how different users perceive the time intervals after the last interaction, we further design a time-interval aware personalized projection operator on the foundation of temporal point processes (TPPs) to project user embedding for the next-time share prediction. Extensive experiments on a real-world e-commerce share dataset have demonstrated that our proposed DynShare can achieve better results compared with state-of-the-art baseline methods. And our code is available on the project website: https://github.com/meteor-gif/DynShare.

Deep metric learning has been widely adopted to construct good representations for images and sentences with pair-based loss functions such as contrastive loss and triplet loss. However, these loss functions restrict the effectiveness of learned embedding and face two critical challenges: 1) high bias on account of a large set of uninformative and redundant pairs; 2) low information exploration due to the lack of processing the relation information among pairs. In this paper, we propose the jointly multi-similarity (JMS) loss to address the above challenges with a novel pair weighting strategy that assigns higher weights to the more informative pairs and discards the less informative ones. Specifically, we integrate the relationship information among pairs into a single framework and optimize the JMS loss by considering various information jointly. Furthermore, we extensively compare the JMS loss with other start-of-the-art approaches by conducting multiple experiments on image retrieval and semantic text similarity tasks. The experimental results show that the JMS loss consistently outperforms competitors.

User habit mining plays an important role in user understanding, which is critical for improving a wide range of personalized intelligence services. Recently, some researchers proposed to mine user behavior patterns which characterize the habits of mobile users and account for the associations between user interactions and context captured by mobile devices. However, the existing approaches for mining these behavior patterns are not practical in mobile environments due to limited computing resources on mobile devices. To fulfill this crucial void, we investigate optimizing strategies which can be used for improving the efficiency of behavior pattern mining in terms of computing and memory needs. Specifically, we examine typical optimizing strategies for association rule mining and study the feasibility of applying them to behavior pattern mining, since these two problems are similar in many aspects. Moreover, we develop an efficient algorithm, named BP-Growth, for behavior pattern mining by combining two promising strategies. Finally, experimental results show that BP-Growth outperforms benchmark methods with a significant margin in terms of both computing and memory cost.

With the prevalence of smart mobile devices with multiple sensors, the commercial application of intelligent context-aware services becomes more and more attractive. However, limited by the battery capacity, the energy efficiency of context-sensing is the bottleneck for the success of context-aware applications. Though several previous studies for energy-efficient context-sensing have been reported, none of them can be applied to multiple types of high-energy-consuming sensors. Moreover, applying machine learning technologies to energy-efficient context-sensing is underexplored too. In this article, we propose to leverage machine learning technologies for improving the energy efficiency of multiple high-energy-consuming context sensors by trading off the sensing accuracy. To be specific, we try to infer the status of high-energy-consuming sensors according to the outputs of software-based sensors and the physical sensors that are necessary to work all the time for supporting the basic functions of mobile devices. If the inference indicates the high-energy-consuming sensor is in a stable status, we avoid the unnecessary invocation and instead use the latest invoked value as the estimation. The experimental results on real datasets show that the energy efficiency of GPS sensing and audio-level sensing are significantly improved by the proposed approach while the sensing accuracy is over 90%.