Graph Transformers (GTs) have emerged as a promising graph learning tool, leveraging their all-pair connected property to effectively capture global information. To address the over-smoothing problem in deep GNNs, global attention was initially introduced, eliminating the necessity for using deep GNNs. However, through empirical and theoretical analysis, we verify that the introduced global attention exhibits severe over-smoothing, causing node representations to become indistinguishable due to its inherent low-pass filtering. This effect is even stronger than that observed in GNNs. To mitigate this, we propose PageRank Transformer (ParaFormer), which features a PageRank-enhanced attention module designed to mimic the behavior of deep Transformers. We theoretically and empirically demonstrate that ParaFormer mitigates over-smoothing by functioning as an adaptive-pass filter. Experiments show that ParaFormer achieves consistent performance improvements across both node classification and graph classification tasks on 11 datasets ranging from thousands to millions of nodes, validating its efficacy. The supplementary material, including code and appendix, can be found in https://github.com/chaohaoyuan/ParaFormer.

Accurate weather forecasting on Subseasonal-to-Seasonal (S2S) timescale is critical to human society such as agriculture planning and extreme weather preparation. Although data-driven models have become alternatives to computationally intensive Numerical Weather Prediction (NWP) systems, existing Transformer-based approaches suffer from biases due to planar projections distorting the spherical geometry and inadequate handling of vector-scalar variable interactions (e.g., wind velocity vs. temperature). To address these limitations, we propose a graph-based Equivariant and Invariant Message Passing (EIMP) framework that directly processes spherical grid data. It maintains SO(3) equivariant embeddings for vector data and SO(3) invariant embeddings for scalar data, which are interacted by a shared invariant message embedding. Guaranteed equivariant and invariant message aggregation functions are proposed to update embeddings under strict symmetry constraints. Extensive experiments on the Earth Reanalysis 5 (ERA5) reanalysis dataset of 41 years demonstrate the proposed model achieves significant improvement over advanced data-driven models and skillful numerical ECMWF systems. Additionally, we empirically show that EIMP demonstrates geometrically superior predictions and conduct ablation studies to validate the efficacy of its design.

The LifeTime Value (LTV) prediction, which endeavors to forecast the cumulative purchase contribution of a user to a particular item, remains a vital challenge that advertisers are keen to resolve. A precise LTV prediction system enhances the alignment of user interests with meticulously designed advertisements, thereby generating substantial profits for advertisers. Nonetheless, this issue is complicated by the paucity of data typically observed in real-world advertising scenarios. The purchase rate among registered users is often as critically low as 0.1%, resulting in a dataset where the majority of users make only several purchases. Consequently, there is insufficient supervisory signal for effectively training the LTV prediction model. An additional challenge emerges from the interdependencies among tasks with high correlation. It is a common practice to estimate a user's contribution to a game over a specified temporal interval. Varying the lengths of these intervals corresponds to distinct predictive tasks, which are highly correlated. For instance, predictions over a 7-day period are heavily reliant on forecasts made over a 3-day period, where exceptional cases can adversely affect the accuracy of both tasks. In order to comprehensively address the aforementioned challenges, we introduce an innovative framework denoted as Graph-Represented Pareto-Optimal LifeTime Value prediction (GRePO-LTV). Graph representation learning is initially employed to address the issue of data scarcity. Subsequently, Pareto-Optimization is utilized to manage the interdependence of prediction tasks. Our method is evaluated using a proprietary offline mini-game recommendation dataset in conjunction with an online A/B test. The implementation of our method results in a significant enhancement within the offline dataset. Moreover, the A/B test demonstrates encouraging outcomes, increasing average Gross Merchandise Value (GMV) by 8.4%.

Information diffusion prediction is a vital component for a wide range of social applications, including viral marketing identification and precise recommendation. Prior methods focus on modeling contextual information from a single cascade, ignoring rich collaborative information behind historical interactions across various cascades and future data within the cascade. Leveraging such interactions can substantially enhance diffusion prediction performance but presents two major challenges: (1) user intents are usually entangled behind historical interactions; and (2) utilizing future data may introduce severe training-inference discrepancies. We present MIM, a novel information diffusion model merging multi-scale interactions for improving user intent learning and behavior retrieval. Specifically, we convert cascades and social relations into multi-channel hypergraphs, where each channel depicts a common fine-grained user intent behind historical interactions across cascades. By aggregating embeddings learned through multiple channels, we obtain comprehensive intent representations. Second, we decouple past- and future-level temporal influences within a cascade via a dual temporal network. Then we implement past-future knowledge transferring to enhance the knowledge learnt from the dual network via hierarchical knowledge distillation. Extensive experiments conducted on four datasets demonstrate that MIM significantly outperforms various benchmarks.

Incorporating Euclidean symmetries (e.g. rotation equivariance) as inductive biases into graph neural networks has improved their generalization ability and data efficiency in unbounded physical dynamics modeling. However, in various scientific and engineering applications, the symmetries of dynamics are frequently discrete due to the boundary conditions. Thus, existing GNNs either over-look necessary symmetry, resulting in suboptimal representation ability, or impose excessive equivariance, which fails to generalize to unobserved symmetric dynamics. In this work, we propose a general Discrete Equivariant Graph Neural Network (DEGNN) that guarantees equivariance to a given discrete point group. Specifically, we show that such discrete equivariant message passing could be constructed by transforming geometric features into permutation-invariant embeddings. Through relaxing continuous equivariant constraints, DEGNN can employ more geometric feature combinations to approximate unobserved physical object interaction functions. Two implementation approaches of DEGNN are proposed based on ranking or pooling permutation-invariant functions. We apply DEGNN to various physical dynamics, ranging from particle, molecular, crowd to vehicle dynamics. In twenty scenarios, DEGNN significantly outperforms existing state-of-the-art approaches. Moreover, we show that DEGNN is data efficient, learning with less data, and can generalize across scenarios such as unobserved orientation.

Anomaly detection (AD) plays a pivotal role in numerous web-based applications, including malware detection, anti-money laundering, device failure detection, and network fault analysis. Most methods, which rely on unsupervised learning, are hard to reach satisfactory detection accuracy due to the lack of labels. Weakly Supervised Anomaly Detection (WSAD) has been introduced with a limited number of labeled anomaly samples to enhance model performance. Nevertheless, it is still challenging for models, trained on an inadequate amount of labeled data, to generalize to unseen anomalies. In this paper, we introduce a novel framework, Knowledge-Data Alignment (KDAlign), to integrate rule knowledge, typically summarized by human experts, to supplement the limited labeled data. Specifically, we transpose these rules into the knowledge space and subsequently recast the incorporation of knowledge as the alignment of knowledge and data. To facilitate this alignment, we employ the Optimal Transport (OT) technique. We then incorporate the OT distance as an additional loss term to the original objective function of WSAD methodologies. Comprehensive experimental results on five real-world datasets demonstrate that our proposed KDAlign framework markedly surpasses its state-of-the-art counterparts, achieving superior performance across various anomaly types. Our codes are released at https://github.com/cshhzhao/KDAlign.

Personalized recommendation is becoming increasingly important in online information systems in the current era of information explosion. In real-world scenarios, when a user considers which items to consume, the decision choice may be affected by her friends. For example, she may ask her friends for suggestions or be attracted by products purchased by one friend. As such, to provide satisfactory recommendation service, it is important to account for the evidence in social relations when they are available to use. Several prior efforts have been made to leverage social relations to build the recommender system and verified their utility. However, most existing methods, such as the well-known TrustSVD, leverage only first-order social relations, i.e., the direct neighbors that are connected to the target user. The high-order social relations, e.g., the friends of friends, which are very informative to reveal user preference, have been largely ignored.

Traditional Click-Through Rate (CTR) prediction models are usually trained and deployed in a single scenario. However, large-scale commercial platforms usually contain multiple recommendation scenarios, the traffic characteristics of which may be significantly different. Recent studies have proved that learning a unified model to serve multiple scenarios is effective in improving the overall performance. However, most existing approaches suffer from various limitations respectively, such as insufficient distinction modeling, inefficiency with the increase of scenarios, and lack of interpretability. More importantly, as far as we know, none of existing Multi-Scenario Modeling approaches takes explicit feature interaction into consideration when modeling scenario distinctions, which limits the expressive power of the network and thus impairs the performance. In this paper, we propose a novel Scenario-Adaptive Feature Interaction framework named SATrans, which models scenario discrepancy as the distinction of patterns in feature correlations. Specifically, SATrans is built on a Transformer architecture to learn high-order feature interaction and involves the scenario information in the modeling of self-attention to capture distribution shifts across scenarios. We provide various implementations of our framework to boost the performance, and experiments on both public and industrial datasets show that SATrans 1) significantly outperforms existing state-of-the-art approaches for prediction, 2) is parameter-efficient as the space complexity grows marginally with the increase of scenarios, 3) offers good interpretability in both instance-level and scenario-level. We have deployed the model in WeChat Official Account Platform and have seen more than 2.84% online CTR increase on average in three major scenarios.

The prevalence of online social network makes it compulsory to study how social relations affect user choice. However, most existing methods leverage only first-order social relations, that is, the direct neighbors that are connected to the target user. The high-order social relations, e.g., the friends of friends, which are very informative to reveal user preference, have been largely ignored. In this work, we focus on modeling the indirect influence from the high-order neighbors in social networks to improve the performance of item recommendation. Distinct from mainstream social recommenders that regularize the model learning with social relations, we instead propose to directly factor social relations in the predictive model, aiming at learning better user embeddings to improve recommendation. To address the challenge that high-order neighbors increase dramatically with the order size, we propose to recursively “propagate” embeddings along the social network, effectively injecting the influence of high-order neighbors into user representation. We conduct experiments on two real datasets of Yelp and Douban to verify ourHigh-Order Social Recommender(HOSR) model. Empirical results show that our HOSR significantly outperforms recent graph regularization-based recommenders NSCR and IF-BPR$^+$, and graph convolutional network-based social influence prediction model DeepInf, achieving new state-of-the-arts of the task.

The sequential recommendation has been widely used to predict users' preferences in the near future by utilizing their dynamic interactions with items. However, existing methods only consider single-typed interactions (e.g., purchase), ignoring the rich heterogeneous information such as multi-typed interactions (e.g., click, purchase) and item attributes (e.g, category), which leads to a suboptimal model. We can integrate this rich information by introducing Dynamic Heterogeneous Information Networks (DHINs). Our solution contains three special designs: 1) Static Initialization; 2) Heterogeneous User Memory Network; 3) Two-level attention mechanism. Extensive experiments conducted on two real-world datasets show that our model outperforms other state-of-the-art solutions. Furthermore, we provide some insights into parameter settings and model interpretability.

Heterogeneous information networks consist of multiple types of edges and nodes, which have a strong ability to represent the rich semantics underpinning network structures. Recently, the dynamics of networks has been studied in many tasks such as social media analysis and recommender systems. However, existing methods mainly focus on the static networks or dynamic homogeneous networks, which are incapable or inefficient in modeling dynamic heterogeneous information networks. In this paper, we propose a method named Dynamic Heterogeneous Information Network Embedding (DyHINE), which can update embeddings when the network evolves. The method contains two key designs: (1) A dynamic time-series embedding module which employs a hierarchical attention mechanism to aggregate neighbor features and temporal random walks to capture dynamic interactions; (2) An online real-time updating module which efficiently updates the computed embeddings via a dynamic operator. Experiments on three real-world datasets demonstrate the effectiveness of our model compared with state-of-the-art methods on the task of temporal link prediction.

Factorization machines (FMs) have been widely adopted to model the discrete feature interactions in recommender systems. Despite their great success, currently there is no study of their robustness to discrete adversarial perturbations. Whether modifying a certain number of the discrete input features has a dramatic effect on the FM's prediction? Although there exist robust training methods for FMs, they neglect the discrete property of input features and lack of an effective mechanism to verify the model robustness.

Heterogeneous Information Network(HIN) has been employed in recommender system to represent heterogeneous types of data, and meta path has been proposed to capture semantic relationship among objects. When applying HIN to the recommendation, there are two problems: how to extract features from meta paths and how to properly fuse these features to further improve recommendations. Some recent work has employed deep neural network to learn user and item representation, and attention mechanism has been explored to integrate information for recommendation. Inspired by these work, in this paper, we propose Heterogeneous Neural Attentive Factorization Machine(HNAFM) to solve above problems. Specifically, we first calculate the commuting matrices based on meta paths and use multilayer perceptrons to learn user and item features. A hierarchical attention mechanism is employed to find the meta path that best describes user's preference and item's property. Comprehensive experiments based on real-world datasets demonstrate that the proposed HNAFM significantly outperforms state-of-the-art rating prediction methods.