When testing data and training data come from different distributions, deep neural networks (DNNs) will face significant safety risks in practical applications. Therefore, out-of-distribution (OOD) detection techniques, which can identify OOD samples at test time and alert the system, are urgently needed. Existing graph OOD detection methods usually characterize fine-grained in-distribution (ID) patterns from multiple perspectives, and train end-to-end graph neural networks (GNNs) for prediction. However, due to the unavailability of OOD data during training, the absence of explicit supervision signals could lead to sub-optimal performance of end-to-end encoders. To address this issue, we follow thepre-training+promptingparadigm to utilize pre-trained GNN encoders, and propose Disentangled Graph Prompting (DGP), to capture fine-grained ID patterns with the help of ID graph labels. Specifically, we design two prompt generators that respectively generate class-specific and class-agnostic prompt graphs by modifying the edge weights of an input graph. We also design several effective losses to train the prompt generators and prevent trivial solutions. We conduct extensive experiments on ten datasets to demonstrate the superiority of our proposed DGP, which achieves a relative AUC improvement of 3.63% over the best graph OOD detection baseline. Ablation studies and hyper-parameter experiments further show the effectiveness of DGP.

Dynamic graphs have attracted increasing attention due to their ability to model complex and evolving relationships in real-world scenarios. Traditional approaches typically pre-train models using dynamic link prediction and directly apply the resulting node temporal embeddings to specific downstream tasks. However, the significant differences among downstream tasks often lead to performance degradation, especially under few-shot settings. Prompt tuning has emerged as an effective solution to this problem. Existing prompting methods are often strongly coupled with specific model architectures or pretraining tasks, which makes it difficult to adapt to recent or future model designs. Moreover, their exclusive focus on modifying node or temporal features while neglecting spatial structural information leads to limited expressiveness and degraded performance. To address these limitations, we propose DDGPrompt, a data-centric prompting framework designed to effectively refine pre-trained node embeddings at the input data level, enabling better adaptability to diverse downstream tasks. We first define a unified node expression feature matrix that aggregates all relevant temporal and structural information of each node, ensuring compatibility with a wide range of dynamic graph models. Then, we introduce three prompt matrices (temporal bias, edge weight, and feature mask) to adjust the feature matrix completely, achieving task-specific adaptation of node embeddings. We evaluate DDGPrompt under a strict few-shot setting on four public dynamic graph datasets. Experimental results demonstrate that our method significantly outperforms traditional methods and prompting approaches in scenarios with limited labels and cold-start conditions.

Protein-protein Interaction (PPI) prediction, which aims to identify the interactions between proteins within a biological system, is an important problem in understanding disease mechanisms and drug discovery. Recently, Equivariant Graph Neural Networks (E3-GNNs) are advanced computational models that provide a powerful solution for accurately predicting PPIs by preserving the geometric integrity of protein interactions. However, most E3-GNNs model protein interactions at the residue level, potentially neglecting critical atomic details and side-chain conformations. In this paper, we propose a novel model, MEANT, designed to adaptively extract atom-level geometric information from varying numbers of atoms within different residues for PPI prediction. Specifically, we define a full-atom graph that contains atomic geometry and guides the message passing under the structure of residues. We also design a geometric relation extractor to integrate geometric information from different residues and adaptively handle variations in the number of atoms within each residue. Finally, we adopt the attention mechanism to update the residue representation and the atomic coordinates within a residue. Experimental results show that our proposed model, MEANT, significantly outperforms state-of-the-art methods on three typical PPI prediction tasks. Our code and data are available on GitHub at https://github.com/BUPT-GAMMA/MEANT.

Large language models (LLMs) have demonstrated their strong capabilities in various domains, and have been recently integrated for graph analysis as graph language models (GLMs). With LLMs as the predictor, some GLMs can interpret unseen tasks described by natural language, and learn from a few examples in the prompts without parameter tuning, known as in-context learning (ICL). Another subset of GLMs utilizes abundant training labels to enhance model performance, known as instruction tuning. However, we argue that ICL on graphs has effectiveness issues due to fixed parameters and efficiency issues due to long context. Meanwhile, the large amount of labeled data required for instruction tuning can be difficult to obtain in real-world scenarios. To this end, we aim to introduce an extra parameter adaptation stage that can efficiently tailor GLMs to an unseen graph and task with only a few labeled examples, in exchange for better prediction accuracy and faster inference speed. For implementation, in this paper we propose GraphLAMA method, with its model backbone and learning schemes specialized for efficient tuning and inference. Specifically, for the model backbone, we use a graph neural network (GNN) with several well-designed components (e.g., hop encodings, gating modules) to transform nodes into the representation space of LLM tokens. Task instructions can then be represented as a mixture of node and language tokens. In the pre-training stage, all model parameters except for the LLM will be trained with different tasks (i.e., node matching, node classification, and link prediction) to capture general knowledge. In the adaptation stage, only a few pre-trained parameters will be updated based on few-shot examples. Extensive experiments on few/zero-shot node classification and summary generation show that our proposed GraphLAMA achieves state-of-the-art (SOTA) performance with 4.91% absolute improvement in accuracy. Compared with ICL, our inference speed can be 10 times faster under 5-shot setting. Our code is available on GitHub at https://github.com/BUPT-GAMMA/GraphLAMA.

Graph-based methods have proven effective in financial fraud detection by modeling relationships between entities, yet they often fail to leverage the rich textual information present in real-world data. With the ability to understand semantic information, large language models (LLMs) offer a promising solution to enhance fraud detection by incorporating textual data, such as user profiles and transaction descriptions. However, integrating LLMs with graph-based methods introduces two key challenges: (1) the neighborhood camouflage problem, where fraudulent nodes disguise themselves within normal network structures, and (2) the input size constraints of LLMs, making it difficult to process large, complex graphs with extensive textual data. In this paper, we propose a novel framework, Fraud Detection with LLM-enhanced Graph Neural Networks (FLAG), to address these challenges. FLAG integrates LLMs with graph-based fraud detection by introducing two main modules: semantic similarity neighbor sampling, which reduces the input size and further alleviates the influence of camouflaged neighbors by selecting neighbors having high semantic similarity with the target nodes, and LLM-based node enhancement, which extracts discriminative textual features by LLM to enhance node robustness against camouflaged neighbors. To further improve the model, we design a fine-tuning approach that enables the LLM to extract discriminative text more closely related to the node labels, enhancing the model's ability to differentiate between fraudulent and normal nodes. Extensive experiments on public datasets highlight the superiority of FLAG, showing average improvements of 3.14% in F1-macro and 6.97% in AUC. Furthermore, we have deployed FLAG in Alipay's credit risk assessment system and evaluated its performance on a real-world dataset. The results indicate a 0.9% improvement in the KS criterion, further underscoring FLAG's effectiveness.

In real-world applications, graph data has garnered significant attention for its representation and analysis using Graph Neural Networks. Recent advancements have led to the development of Graph Foundation Models (GFMs), which aim to enhance cross-domain and cross-task generalization ability. Despite promising results from GFMs, a lack of standardized evaluation processes hinders comparative analysis and cross-domain applicability. To address this gap, we propose GFMBench, an open-source pipeline that standardizes the training, evaluation, and deployment of GFMs across diverse real-world graph applications. GFMBench integrates state-of-the-art GFMs and datasets, providing a modular design for comprehensive support across data preprocessing, model training, and evaluation. The pipeline includes a robust evaluation framework for benchmarking GFM generalization ability, encompassing supervised learning, cross-domain zero-shot and few-shot learning, and in-context learning. To validate the usability of GFMs, we deploy them on the Open Academic Graph, enabling applications such as topic search and author recommendation. This work provides a unified benchmark for GFMs, enabling deeper insights into their generalization ability across various graph tasks and domains. We further open-source GFMBench https://github.com/BUPT-GAMMA/ggfm and related documents https://ggfm.readthedocs.io/en/latest/.

Recommender systems (RSs) are designed to retrieve candidate items a user might be interested in from a large pool, with a typical approach being the use of graph neural networks (GNNs) to capture high-order interaction relationships. As large language models (LLMs) have demonstrated remarkable success across various domains, researchers are exploring ways to apply their capabilities for improving recommendation performance. However, existing work limits the use of LLMs to either re-ranking recommendation results of traditional RSs or pre-processing the datasets as data augmenters. Both lines of work failed to explore LLMs' capabilities during the filtering process of candidate items, which may lead to suboptimal performance. Instead, we propose to leverage LLMs' reasoning abilities during the candidate filtering process, and introduce Chain Of Retrieval ON grAphs (CORONA) to progressively narrow down the range of candidate items on interaction graphs with the help of LLMs: (1) First, LLM performs preference reasoning based on user profiles, with the response serving as a query to extract relevant users and items from the interaction graph as preference-assisted retrieval ; (2) Then, using the information retrieved in the previous step along with the purchase history of target user, LLM conducts intent reasoning to help refine an even smaller interaction subgraph as intent-assisted retrieval ; (3) Finally, we employ a GNN to capture high-order collaborative filtering information from the extracted subgraph, performing GNN-enhanced retrieval to generate the final recommendation results. The proposed framework leverages the reasoning capabilities of LLMs during the retrieval process, while seamlessly integrating GNNs to enhance overall recommendation performance. Extensive experiments on various datasets and settings demonstrate that our proposed CORONA achieves state-of-the-art (SOTA) performance with an 18.6% relative improvement in recall and an 18.4% relative improvement in NDCG on average. Our code is available on GitHub at https://github.com/BUPT-GAMMA/CORONA.

With the expansion of business scenarios, real recommender systems are facing challenges in dealing with the constantly emerging new tasks in multi-task learning frameworks. In this article, we attempt to improve the generalization ability of multi-task recommendations when dealing with new tasks. A novel two-stage prompt-tuning MTL framework (MPT-Rec) is proposed to address task irrelevance and training efficiency problems in multi-task recommender systems. Specifically, we disentangle the task-specific and task-sharing information in the multi-task pre-training stage and then use task-aware prompts to transfer knowledge from other tasks to the new task effectively. By freezing parameters in the pre-training tasks, MPT-Rec solves the negative impacts that may be brought by the new task and greatly reduces the training costs. Extensive experiments on three real-world datasets show the effectiveness of our proposed multi-task learning framework. MPT-Rec achieves the best performance compared to the SOTA multi-task learning method on three real-world datasets. Besides, it maintains comparable model performance but vastly improves the training efficiency (i.e., with up to 10% parameters in the full-training way) in the new task learning. Our code is publicly available at https://github.com/BAI-LAB/MPT-Rec .

Graph neural networks (GNNs) have revolutionized molecule representation learning by modeling molecules as graphs, with atoms represented as nodes and chemical bonds as edges. Despite their progress, they struggle with out-of-distribution scenarios, such as changes in size or scaffold of molecules with identical properties. Some studies attempt to mitigate this issue through graph invariant learning, which penalizes prediction variance across environments to learn invariant representations. But in the realm of molecules, core functional groups forming privileged substructures dominate molecular properties and remain invariant across distribution shifts. This highlights the need for integrating this prior knowledge and ensuring the environment split compatible with molecule invariant learning. To bridge this gap, we propose a novel framework named MILI. Specifically, we first formalize molecule invariant learning based on privileged substructure identification and introduce substructure invariance constraint. Building on this foundation, we theoretically establish two criteria for environment splits conducive to molecule invariant learning. Inspired by these criteria, we develop a dual-head graph neural network. A shared identifier identifies privileged substructures, while environment and task heads generate predictions based on variant and privileged substructures. Through the interaction of two heads, the environments are split and optimized to meet our criteria. The unified MILI guarantees that molecule invariant learning and environment split achieve mutual enhancement from theoretical analysis and network design. Extensive experiments across eight benchmarks validate the effectiveness of MILI compared to state-of-the-art baselines.

Graph neural networks (GNNs) have gained popularity in modeling various complex networks, e.g., social network and webpage network. Despite the promising accuracy, the confidences of GNNs are shown to be miscalibrated, indicating limited awareness of prediction uncertainty and harming the reliability of model decisions. Existing calibration methods primarily focus on improving GNN models, e.g., adding regularization during training or introducing temperature scaling after training. In this paper, we argue that the miscalibration of GNNs may stem from the graph data and can be alleviated through topology modification. To support this motivation, we conduct data observations by examining the impacts ofdecisive andhomophilic edges on calibration performance, where decisive edges play a critical role in GNN predictions and homophilic edges connect nodes of the same class. By assigning larger weights to these edges in the adjacency matrix, we observe an improvement in calibration performance without sacrificing classification accuracy. This suggests the potential of a data-centric approach for calibrating GNNs. Motivated by our observations, we propose Data-centric Graph Calibration (DCGC), which uses two edge weighting modules to adjust the input graph for GNN calibration. The first module learns the weights of decisive edges by parameterizing the adjacency matrix and enabling backpropagation of the prediction loss to edge weights. This emphasizes critical edges that fit the prediction needs. The second module computes weights for homophilic edges based on predicted label distributions, assigning larger weights to edges with stronger homophily. These modifications operate at the data level and can be easily integrated with temperature scaling-based methods for better calibration. Experimental results on 8 benchmark datasets demonstrate that DCGC achieves state-of-the-art calibration performance, with an average relative improvement of 36.4% in ECE, while maintaining or even slightly improving classification accuracy. Ablation studies and hyper-parameter analysis further validate the effectiveness and robustness of our proposed method DCGC. Code and data are available at https://github.com/BUPT-GAMMA/DCGC.

Large language models (LLMs) like ChatGPT, exhibit powerful zero-shot and instruction-following capabilities, have catalyzed a revolutionary transformation across diverse fields, especially for open-ended tasks. While the idea is less explored in the graph domain, despite the availability of numerous powerful graph models (GMs), they are restricted to tasks in a pre-defined form. Although several methods applying LLMs to graphs have been proposed, they fail to simultaneously handle the pre-defined and open-ended tasks, with LLM as a node feature enhancer or as a standalone predictor. To break this dilemma, we propose to bridge the pretrained GM and LLM by a Translator, named GraphTranslator, aiming to leverage GM to handle the pre-defined tasks effectively and utilize the extended interface of LLMs to offer various open-ended tasks for GM. To train such Translator, we propose a Producer capable of constructing the graph-text alignment data along node information, neighbor information and model information. By translating node representation into tokens, GraphTranslator empowers an LLM to make predictions based on language instructions, providing a unified perspective for both pre-defined and open-ended tasks. Extensive results demonstrate the effectiveness of our proposed GraphTranslator on zero-shot node classification. The graph question answering experiments reveal our GraphTranslator potential across a broad spectrum of open-ended tasks through language instructions. Our code is available at: https://github.com/alibaba/GraphTranslator

Pre-trained graph models (PGMs) aim to capture transferable inherent structural properties and apply them to different downstream tasks. Similar to pre-trained language models, PGMs also inherit biases from human society, resulting in discriminatory behavior in downstream applications. The debiasing process of existing fair methods is generally coupled with parameter optimization of GNNs. However, different downstream tasks may be associated with different sensitive attributes in reality, directly employing existing methods to improve the fairness of PGMs is inflexible and inefficient. Moreover, most of them lack a theoretical guarantee, i.e., provable lower bounds on the fairness of model predictions, which directly provides assurance in a practical scenario. To overcome these limitations, we propose a novel adapter-tuning framework that endows pre-trained Graph models with Provable fAiRness (called GraphPAR). GraphPAR freezes the parameters of PGMs and trains a parameter-efficient adapter to flexibly improve the fairness of PGMs in downstream tasks. Specifically, we design a sensitive semantic augmenter on node representations, to extend the node representations with different sensitive attribute semantics for each node. The extended representations will be used to further train an adapter, to prevent the propagation of sensitive attribute semantics from PGMs to task predictions. Furthermore, with GraphPAR, we quantify whether the fairness of each node is provable, i.e., predictions are always fair within a certain range of sensitive attribute semantics. Experimental evaluations on real-world datasets demonstrate that GraphPAR achieves state-of-the-art prediction performance and fairness on node classification task. Furthermore, based on our GraphPAR, around 90% nodes have provable fairness.

The explosive growth of cyber attacks today, such as malware, spam, and intrusions, has caused severe consequences on society. Securing cyberspace has become a great concern for organizations and governments. Traditional machine learning based methods are extensively used in detecting cyber threats, but they hardly model the correlations between real-world cyber entities. In recent years, with the proliferation of graph mining techniques, many researchers have investigated these techniques for capturing correlations between cyber entities and achieving high performance. It is imperative to summarize existing graph-based cybersecurity solutions to provide a guide for future studies. Therefore, as a key contribution of this work, we provide a comprehensive review of graph mining for cybersecurity, including an overview of cybersecurity tasks, the typical graph mining techniques, and the general process of applying them to cybersecurity, as well as various solutions for different cybersecurity tasks. For each task, we probe into relevant methods and highlight the graph types, graph approaches, and task levels in their modeling. Furthermore, we collect open datasets and toolkits for graph-based cybersecurity. Finally, we present an outlook on the potential directions of this field for future research.

Heterogeneous graph neural network (HGNN) has shown superior performance and attracted considerable research interest. However, HGNN inherits the limitation of expressive power from GNN via learning$individual$node embeddings based on their structural neighbors, largely ignoring the potential correlations between nodes and leading to sub-optimal performance.How to establish correlations among multiple node embeddings and improve the expressive power of HGNN is still an open problem.To solve the above problem, we propose a simple and effective technique called heterogeneous distance encoding (HDE) to fundamentally improve the expressive power of HGNN. Specifically, we define heterogeneous shortest path distance to describe the relative distance between nodes, and then jointly encode such distances for multiple nodes of interest to establish their correlation. By simply injecting the encoded correlation into the neighbor aggregating process, we can learn more expressive heterogeneous graph representations for downstream tasks. More importantly, the proposed HDE relies only on the graph structure and ensures the inductive ability of HGNN. We also propose an efficient HDE algorithm that can significantly reduce the computational overhead. Significant improvements on both transductive and inductive tasks over four real-world graphs demonstrate the effectiveness of HDE in improving the expressive power of HGNN.

In recent years, Heterogeneous Graph Neural Networks (HGNNs) have gained increasing attention due to their excellent performance in applications. However, the lack of high-quality benchmarks in new fields has become a critical limitation for developing and applying HGNNs. To accommodate the urgent need for emerging fields and the advancement of HGNNs, we present two large-scale, real-world, and challenging heterogeneous graph datasets from real scenarios: risk commodity detection and takeout recommendation. Meanwhile, we establish standard benchmark interfaces that provide over 40 heterogeneous graph datasets. We provide initial data split, unified evaluation metrics, and baseline results for future work, making it fair and handy to explore state-of-the-art HGNNs. Our interfaces also offer a comprehensive toolkit to research the characteristics of graph datasets. The above new datasets are publicly available on https://zenodo.org/communities/hgd, and the interface codes are available at https://github.com/BUPT-GAMMA/hgbi.

Out-of-distribution (OOD) detection, which aims to identify OOD samples from in-distribution (ID) ones in test time, has become an essential problem in machine learning. However, existing works are mostly conducted on Euclidean data, and the problem in graph-structured data remains under-explored. Several recent works begin to study graph OOD detection, but they all need to train a graph neural network (GNN) from scratch with high computational cost. In this work, we make the first attempt to endow a well-trained GNN with the OOD detection ability without modifying its parameters. To this end, we design a post-hoc framework with Adaptive Amplifier for Graph OOD Detection, named AAGOD, concentrating on data-centric manipulation. The insight of AAGOD is to superimpose a parameterized amplifier matrix on the adjacency matrix of each original input graph. The amplifier can be seen as prompts and is expected to emphasize the key patterns helpful for graph OOD detection, thereby enlarging the gap between OOD and ID graphs. Then well-trained GNNs can be reused to encode the amplified graphs into vector representations, and pre-defined scoring functions can further convert the representations into detection scores. Specifically, we design a Learnable Amplifier Generator (LAG) to customize amplifiers for different graphs, and propose a Regularized Learning Strategy (RLS) to train parameters with no OOD data required. Experiment results show that AAGOD can be applied on various GNNs to enable the OOD detection ability. Compared with the state-of-the-art baseline in graph OOD detection, on average AAGOD has 6.21% relative enhancement in AUC and a 34 times faster training speed. Code and data are available at https://github.com/BUPT-GAMMA/AAGOD.

Abnormal event detection, which refers to mining unusual interactions among involved entities, plays an important role in many real applications. Previous works mostly oversimplify this task as detecting abnormal pair-wise interactions. However, real-world events may contain multi-typed attributed entities and complex interactions among them, which forms an Attributed Heterogeneous Information Network (AHIN). With the boom of social networks, abnormal event detection in AHIN has become an important, but seldom explored task. In this paper, we firstly study the unsupervised abnormal event detection problem in AHIN. The events are considered as star-schema instances of AHIN and are further modeled by hypergraphs. A novel hypergraph contrastive learning method, named AEHCL, is proposed to fully capture abnormal event patterns. AEHCL designs the intra-event and inter-event contrastive modules to exploit self-supervised AHIN information. The intra-event contrastive module captures the pair-wise and multivariate interaction anomalies within an event, and the inter-event module captures the contextual anomalies among events. These two modules collaboratively boost the performance of each other and improve the detection results. During the testing phase, a contrastive learning-based abnormal event score function is further proposed to measure the abnormality degree of events. Extensive experiments on three datasets in different scenarios demonstrate the effectiveness of AEHCL, and the results improve state-of-the-art baselines up to 12.0% in Average Precision (AP) and 4.6% in Area Under Curve (AUC) respectively.

Graph Neural Networks (GNNs) have shown their superiority in modeling graph-structured data, and gained much attention over the last five years. Though traditional deep learning frameworks such as TensorFlow and PyTorch provide convenient tools for implementing neural network algorithms, they do not support the key operations of GNNs well, e.g., the message passing computation based on sparse matrices. To address this issue, GNN libraries such as PyG are proposed by introducing rich Application Programming Interfaces (APIs) specialized for GNNs. However, most current GNN libraries only support a specific deep learning framework as the backend, e.g., PyG is tied up with PyTorch. In practice, users usually need to combine GNNs with other neural network components, which may come from their co-workers or open-source codes with different deep-learning backends. Consequently, users have to be familiar with various GNN libraries, and rewrite their GNNs with corresponding APIs. To provide a more convenient user experience, we present Gamma Graph Library (GammaGL), a GNN library that supports multiple deep learning frameworks as backends. GammaGL uses a framework-agnostic design that allows users to easily switch between deep learning backends on top of existing components with a single line of code change. Following the tensor-centric design idea, GammaGL splits the graph data into several key tensors, and abstracts GNN computational processes (such as message passing and graph mini-batch operations) into a few key functions. We develop many efficient operators in GammaGL for acceleration. So far, GammaGL has provided more than 40 GNN examples that can be applied to a variety of downstream tasks. GammaGL also provides tools for heterogeneous graph neural networks and recommendations to facilitate research in related fields. We present the performance of models implemented by GammaGL and the time consumption of our optimized operators to show the efficiency. Our library is available at https://github.com/BUPT-GAMMA/GammaGL.

By jointly modeling user-item interactions and knowledge graph (KG) information, KG-based recommender systems have shown their superiority in alleviating data sparsity and cold start problems. Recently, graph neural networks (GNNs) have been widely used in KG-based recommendation, owing to the strong ability of capturing high-order structural information. However, we argue that existing GNN-based methods have the following two limitations. Interaction domination: the supervision signal of user-item interaction will dominate the model training, and thus the information of KG is barely encoded in learned item representations; Knowledge overload: KG contains much recommendation-irrelevant information, and such noise would be enlarged during the message aggregation of GNNs. The above limitations prevent existing methods to fully utilize the valuable information lying in KG. In this paper, we propose a novel algorithm named Knowledge-Adaptive Contrastive Learning (KACL) to address these challenges. Specifically, we first generate data augmentations from user-item interaction view and KG view separately, and perform contrastive learning across the two views. Our design of contrastive loss will force the item representations to encode information shared by both views, thereby alleviating the interaction domination issue. Moreover, we introduce two learnable view generators to adaptively remove task-irrelevant edges during data augmentation, and help tolerate the noises brought by knowledge overload. Experimental results on three public benchmarks demonstrate that KACL can significantly improve the performance on top-K recommendation compared with state-of-the-art methods.

Graph Neural Networks (GNNs) can effectively capture both the topology and attribute information of a graph, and have been extensively studied in many domains. Recently, there is an emerging trend that equips GNNs with knowledge distillation for better efficiency or effectiveness. However, to the best of our knowledge, existing knowledge distillation methods applied on GNNs all employed predefined distillation processes, which are controlled by several hyper-parameters without any supervision from the performance of distilled models. Such isolation between distillation and evaluation would lead to suboptimal results. In this work, we aim to propose a general knowledge distillation framework that can be applied on any pretrained GNN models to further improve their performance. To address the isolation problem, we propose to parameterize and learn distillation processes suitable for distilling GNNs. Specifically, instead of introducing a unified temperature hyper-parameter as most previous work did, we will learn node-specific distillation temperatures towards better performance of distilled models. We first parameterize each node's temperature by a function of its neighborhood's encodings and predictions, and then design a novel iterative learning process for model distilling and temperature learning. We also introduce a scalable variant of our method to accelerate model training. Experimental results on five benchmark datasets show that our proposed framework can be applied on five popular GNN models and consistently improve their prediction accuracies with 3.12% relative enhancement on average. Besides, the scalable variant enables 8 times faster training speed at the cost of 1% prediction accuracy.

Recent years have witnessed the success of heterogeneous graph neural networks (HGNNs) in modeling heterogeneous information networks (HINs). In this paper, we focus on the benchmark task of HGNNs, i.e., node classification, and empirically find that typical HGNNs are not good at predicting the label of a test node whose receptive field (1) has few training nodes from the same category or (2) has multiple training nodes from different categories. A possible explanation is that their message passing mechanisms may involve noises from different categories, and cannot fully explore task-specific knowledge such as the label dependency between distant nodes. Therefore, instead of introducing a new HGNN model, we propose a general post-training framework that can be applied on any pretrained HGNNs to further inject task-specific knowledge and enhance their prediction performance. Specifically, we first design an auxiliary system that estimates node labels based on (1) a global inference module of multi-channel label propagation and (2) a local inference module of network schema-aware prediction. The mechanism of our auxiliary system can complement the pretrained HGNNs by providing extra task-specific knowledge. During the post-training process, we will strengthen both system-level and module-level consistencies to encourage the cooperation between a pretrained HGNN and our auxiliary system. In this way, both systems can learn from each other for better performance. In experiments, we apply our framework to four typical HGNNs. Experimental results on three benchmark datasets show that compared with pretrained HGNNs, our post-training framework can enhance Micro-F1 by a relative improvement of on average. Code, data and appendix are available at https://github.com/GXM1141/HGPF.

Heterogeneous Graph Neural Networks (HGNNs), as a kind of powerful graph representation learning methods on heterogeneous graphs, have attracted increasing attention of many researchers. Although, several existing libraries have supported HGNNs, they just provide the most basic models and operators. Building and benchmarking various downstream tasks on HGNNs is still painful and time consuming with them. In this paper, we will introduce OpenHGNN, an open-source toolkit for HGNNs. OpenHGNN defines a unified and standard pipeline for training and testing, which can allow users to run a model on a specific dataset with just one command line. OpenHGNN has integrated 20+ mainstream HGNNs and 20+ heterogeneous graph datasets, which can be used for various advanced tasks, such as node classification, link prediction, and recommendation. In addition, thanks to the modularized design of OpenHGNN, it can be extended to meet users' customized needs. We also release several novel and useful tools and features, including leaderboard, autoML, design space, and visualization, to provide users with better usage experiences. OpenHGNN is an open-source project, and the source code is available at https://github.com/BUPT-GAMMA/OpenHGNN.

Event argument extraction (EAE) is an important information extraction task, which aims to identify the arguments of an event described in a given text and classify the roles played by them. A key characteristic in realistic EAE data is that the instance numbers of different roles follow an obvious long-tail distribution. However, the training and evaluation paradigms of existing EAE models either prone to neglect the performance on "tail roles'', or change the role instance distribution for model training to an unrealistic uniform distribution. Though some generic methods can alleviate the class imbalance in long-tail datasets, they usually sacrifice the performance of "head classes'' as a trade-off. To address the above issues, we propose to train our model on realistic long-tail EAE datasets, and evaluate the average performance over all roles. Inspired by the Mixture of Experts (MOE), we propose a Routing-Balanced Dual Expert Framework (RBDEF), which divides all roles into "head" and "tail" two scopes and assigns the classifications of head and tail roles to two separate experts. In inference, each encoded instance will be allocated to one of the two experts by a routing mechanism. To reduce routing errors caused by the imbalance of role instances, we design a Balanced Routing Mechanism (BRM), which transfers several head roles to the tail expert to balance the load of routing, and employs a tri-filter routing strategy to reduce the misallocation of the tail expert's instances. To enable an effective learning of tail roles with scarce instances, we devise Target-Specialized Meta Learning (TSML) to train the tail expert. Different from other meta learning algorithms that only search a generic parameter initialization equally applying to infinite tasks, TSML can adaptively adjust its search path to obtain a specialized initialization for the tail expert, thereby expanding the benefits to the learning of tail roles. In experiments, RBDEF significantly outperforms the state-of-the-art EAE models and advanced methods for long-tail data.

With the rapid growth of interaction data, many clustering methods have been proposed to discover interaction patterns as prior knowledge beneficial to downstream tasks. Considering that an interaction can be seen as an action occurring among multiple objects, most existing methods model the objects and their pair-wise relations as nodes and links in graphs. However, they only model and leverage part of the information in real entire interactions, i.e., either decompose the entire interaction into several pair-wise sub-interactions for simplification, or only focus on clustering some specific types of objects, which limits the performance and explainability of clustering. To tackle this issue, we propose to Co-cluster the Interactions via Attentive Hypergraph neural network (CIAH). Particularly, with more comprehensive modeling of interactions by hypergraph, we propose an attentive hypergraph neural network to encode the entire interactions, where an attention mechanism is utilized to select important attributes for explanations. Then, we introduce a salient method to guide the attention to be more consistent with real importance of attributes, namely saliency-based consistency. Moreover, we propose a novel co-clustering method to perform a joint clustering for the representations of interactions and the corresponding distributions of attribute selection, namely cluster-based consistency. Extensive experiments demonstrate that our CIAH significantly outperforms state-of-the-art clustering methods on both public datasets and real industrial datasets.

Heterogeneous Graph Neural Network (HGNN) has been successfully employed in various tasks, but we cannot accurately know the importance of different design dimensions of HGNNs due to diverse architectures and applied scenarios. Besides, in the research community of HGNNs, implementing and evaluating various tasks still need much human effort. To mitigate these issues, we first propose a unified framework covering most HGNNs, consisting of three components: heterogeneous linear transformation, heterogeneous graph transformation, and heterogeneous message passing layer. Then we build a platform Space4HGNN by defining a design space for HGNNs based on the unified framework, which offers modularized components, reproducible implementations, and standardized evaluation for HGNNs. Finally, we conduct experiments to analyze the effect of different designs. With the insights found, we distill a condensed design space and verify its effectiveness.

Dynamic link prediction, which aims at forecasting future edges of a node in a dynamic network, is an important problem in network science and has a wide range of real-world applications. A key property of dynamic networks is that new nodes and links keep coming over time and these new nodes usually have only a few links at their arrivals. However, how to predict future links for these few-shot nodes in a dynamic network has not been well studied. Existing dynamic network representation learning methods were not specialized for few-shot scenarios and thus would lead to suboptimal performances. In this paper, we propose a novel model based on a meta-learning framework, dubbed as MetaDyGNN, for few-shot link prediction in dynamic networks. Specifically, we propose a meta-learner with hierarchical time interval-wise and node-wise adaptions to extract general knowledge behind this problem. We also design a simple and effective dynamic graph neural network (GNN) module to characterize the local structure of each node in meta-learning tasks. As a result, the learned general knowledge serves as model initializations, and can quickly adapt to new nodes with a fine-tuning process on only a few links. Experimental results show that our proposed MetaDyGNN significantly outperforms state-of-the-art methods on three publicly available datasets.

There are plenty of parking spaces in big cities, but we often find nowhere to park. For example, New York has 1.4 million cars and 4.4 million on-street parking spaces, but it is still not easy to find a parking place near our destination, especially during peak hours. The reason is the lack of prediction of parking behavior. If we could provide parking behavior in advance, we can ease this parking problem that affects human well-being. We observe that parking lots have periodic parking patterns, which is an important factor for parking behavior prediction. Unfortunately, existing work ignores such periodic parking patterns in parking behavior prediction, and thus incurs low accuracy. To solve this problem, we propose PewLSTM, a novel periodic weather-aware LSTM model that successfully predicts the parking behavior based on historical records, weather, environments, weekdays, and events. PewLSTM includes a periodic weather-aware LSTM prediction module and an event prediction module, for predicting parking behaviors in regular days and events. PewLSTM is extremely useful for drivers and parking lot owners to improve customer experience. For example, the probability of parking space that will be available soon can be provided even if the parking lot is full. Based on 910,477 real parking records in 904 days from 13 parking lots, PewLSTM yields 93.84% parking prediction accuracy, which is about 30% higher than the state-of-the-art parking behavior prediction method. Additionally, we have analyzed parking behaviors in events like holidays and COVID-19. PewLSTM can handle parking behavior prediction in events and reaches 90.68 percent accuracy.

There is recently a surge in approaches that learn low-dimensional embeddings of nodes in networks. However, for large-scale real-world networks, it’s inefficient for existing approaches to store amounts of parameters in memory and update them edge by edge. With the knowledge that nodes having similar neighborhoods will be close to each other in the embedding space, we propose COSINE (COmpresSIve Network Embedding) algorithm, which reduces the memory footprint and accelerates the training process by parameter sharing among similar nodes. COSINE applies graph partitioning algorithms to networks and builds parameter sharing dependency of nodes based on the results of partitioning. In this way, COSINE injects prior knowledge about high-order structural information into models, which makes network embedding more efficient and effective. COSINE can be applied to anyembedding lookupmethod and learn high-quality embeddings with limited memory and less training time. We conduct experiments on multi-label classification and link prediction, where baselines and our model have the same memory usage. Experimental results show that COSINE improves baselines by up to 23 percent on classification and 25 percent on link prediction. Moreover, the training time of all representation learning methods using COSINE decreases by 30 to 70 percent.

Relation extraction (RE), an important information extraction task, faced the great challenge brought by limited annotation data. To this end, distant supervision was proposed to automatically label RE data, and thus largely increased the number of annotated instances. Unfortunately, lots of noise relation annotations brought by automatic labeling become a new obstacle. Some recent studies have shown that the teacher-student framework of knowledge distillation can alleviate the interference of noise relation annotations via label softening. Nevertheless, we find that they still suffer from two problems: propagation of inaccurate dark knowledge and constraint of a unified distillation temperature . In this article, we propose a simple and effective Multi-instance Dynamic Temperature Distillation (MiDTD) framework, which is model-agnostic and mainly involves two modules: multi-instance target fusion (MiTF) and dynamic temperature regulation (DTR). MiTF combines the teacher’s predictions for multiple sentences with the same entity pair to amend the inaccurate dark knowledge in each student’s target. DTR allocates alterable distillation temperatures to different training instances to enable the softness of most student’s targets to be regulated to a moderate range. In experiments, we construct three concrete MiDTD instantiations with BERT, PCNN, and BiLSTM-based RE models, and the distilled students significantly outperform their teachers and the state-of-the-art (SOTA) methods.

Information diffusion prediction targets on forecasting how information items spread among a set of users. Recently, neural networks have been widely used in modeling information diffusion, owing to the great successes of deep learning. However, in real-world information diffusion scenarios, users are likely to have different behaviors to information items from different topics. Existing neural-based methods failed to model the topic-specific diffusion patterns and dependencies, which have been shown to be useful in conventional non-neural methods. In this paper, we propose Topic-aware Attention Network (TAN) to take advantage of both topic-specific diffusion modeling and deep learning techniques. We jointly model the text content of information items and cascade sequences by incorporating topical context and user/position dependencies into user representations via attention mechanisms. A time-decayed aggregation module is further employed to integrate user representations for cascade representations, which can encode the topic-specific diffusion dependencies independently. Experimental results on diffusion prediction tasks over three realistic cascade datasets show that our model can achieve a relative improvement up to 9% against the best performing baseline in terms of [email protected]

Heterogeneous graphs (HGs), consisting of multiple types of nodes and links, can characterize a variety of real-world complex systems. Recently, heterogeneous graph neural networks (HGNNs), as a powerful graph embedding method to aggregate heterogeneous structure and attribute information, has earned a lot of attention. Despite the ability of HGNNs in capturing rich semantics which reveal different aspects of nodes, they still stay at a coarse-grained level which simply exploits structural characteristics. In fact, rich unstructured text content of nodes also carries latent but more fine-grained semantics arising from multi-facet topic-aware factors, which fundamentally manifest why nodes of different types would connect and form a specific heterogeneous structure. However, little effort has been devoted to factorizing them.

Heterogeneous graph neural network (HGNN) has shown superior performance and attracted considerable research interest. However, HGNN inherits the limitation of representational power from GNN via learning individual node embeddings based on their neighbors, largely ignoring the potential correlations between nodes. In fact, the complex correlation between nodes (e.g., distance) is crucial for many graph mining tasks. How to establish correlations between multiple node embeddings and improve the representational power of HGNN is still an open problem. To solve it, we propose a heterogeneous distance encoding (HDE) technique to fundamentally improve the representational power of HGNN. Specifically, we define heterogeneous shortest path distance to describe the relative distance between nodes, and then jointly encode such distances for multiple nodes of interest to establish their correlation. By simply injecting the encoded correlation into the neighbor aggregating process, we propose a novel distance encoding based heterogeneous graph neural network (called DHN), which is able to learn more expressive heterogeneous graph representations for downstream tasks. More importantly, the proposed DHN relies only on the graph structure and ensures the inductive ability of HGNN. Significant improvements over four real-world graphs demonstrate the representational power of HDE.

With the booming of online social networks in the mobile internet, an emerging recommendation scenario has played a vital role in information acquisition for user, where users are no longer recommended with a single item or item list, but a combination of heterogeneous and diverse objects (called a package, e.g., a package including news, publisher, and friends viewing the news). Different from the conventional recommendation where users are recommended with the item itself, in package recommendation, users would show great interests on the explicitly displayed objects that could have a significant influence on the user behaviors. However, to the best of our knowledge, few effort has been made for package recommendation and existing approaches can hardly model the complex interactions of diverse objects in a package. Thus, in this paper, we make a first study on package recommendation and propose an Intra- and inter-package attention network for Package Recommendation (IPRec). Specifically, for package modeling, an intra-package attention network is put forward to capture the object-level intention of user interacting with the package, while an inter-package attention network acts as a package-level information encoder that captures collaborative features of neighboring packages. In addition, to capture users preference representation, we present a user preference learner equipped with a fine-grained feature aggregation network and coarse-grained package aggregation network. Extensive experiments on three real-world datasets demonstrate that IPRec significantly outperforms the state of the arts. Moreover, the model analysis demonstrates the interpretability of our IPRec and the characteristics of user behaviors. Codes and datasets can be obtained at https://github.com/LeeChenChen/IPRec.

Semi-supervised learning on graphs is an important problem in the machine learning area. In recent years, state-of-the-art classification methods based on graph neural networks (GNNs) have shown their superiority over traditional ones such as label propagation. However, the sophisticated architectures of these neural models will lead to a complex prediction mechanism, which could not make full use of valuable prior knowledge lying in the data, e.g., structurally correlated nodes tend to have the same class. In this paper, we propose a framework based on knowledge distillation to address the above issues. Our framework extracts the knowledge of an arbitrary learned GNN model (teacher model), and injects it into a well-designed student model. The student model is built with two simple prediction mechanisms, i.e., label propagation and feature transformation, which naturally preserves structure-based and feature-based prior knowledge, respectively. In specific, we design the student model as a trainable combination of parameterized label propagation and feature transformation modules. As a result, the learned student can benefit from both prior knowledge and the knowledge in GNN teachers for more effective predictions. Moreover, the learned student model has a more interpretable prediction process than GNNs. We conduct experiments on five public benchmark datasets and employ seven GNN models including GCN, GAT, APPNP, SAGE, SGC, GCNII and GLP as the teacher models. Experimental results show that the learned student model can consistently outperform its corresponding teacher model by on average. Code and data are available at https://github.com/BUPT-GAMMA/CPF

Rumors spread dramatically fast through online social media services, and people are exploring methods to detect rumors automatically. Existing methods typically learn semantic representations of all reposts to a rumor candidate for prediction. However, it is crucial to efficiently detect rumors as early as possible before they cause severe social disruption, which has not been well addressed by previous works. In this paper, we present a novel early rumor detection model, Credible Early Detection (CED). By regarding all reposts to a rumor candidate as a sequence, the proposed model will seek an early point-in-time for making a credible prediction. We conduct experiments on three real-world datasets, and the results demonstrate that our proposed model can remarkably reduce the time span for prediction by more than 85 percent, with better accuracy performance than all state-of-the-art baselines.

The study of information diffusion or cascade has attracted much attention over the last decade. Most related works target on studying cascade-level macroscopic properties such as the final size of a cascade. Existing microscopic cascade models which focus on user-level modeling either make strong assumptions on how a user gets infected by a cascade or limit themselves to a specific scenario where “who infected whom” information is explicitly labeled. The strong assumptions oversimplify the complex diffusion mechanism and prevent these models from better fitting real-world cascade data. Also, the methods which focus on specific scenarios cannot be generalized to a general setting where the diffusion graph is unobserved. To overcome the drawbacks of previous works, we propose a Neural Diffusion Model (NDM) for general microscopic cascade study. NDM makes relaxed assumptions and employs deep learning techniques including attention mechanism and convolutional network for cascade modeling. Both advantages enable our model to go beyond the limitations of previous methods, better fit the diffusion data and generalize to unseen cascades. Experimental results on diffusion identification task over four realistic cascade datasets show that our model can achieve a relative improvement up to 26 percent against the best performing baseline in terms of F1 score.

Attributed graph embedding, which learns vector representations from graph topology and node features, is a challenging task for graph analysis. Recently, methods based on graph convolutional networks (GCNs) have made great progress on this task. However,existing GCN-based methods have three major drawbacks. Firstly,our experiments indicate that the entanglement of graph convolutional filters and weight matrices will harm both the performance and robustness. Secondly, we show that graph convolutional filters in these methods reveal to be special cases of generalized Laplacian smoothing filters, but they do not preserve optimal low-pass characteristics. Finally, the training objectives of existing algorithms are usually recovering the adjacency matrix or feature matrix, which are not always consistent with real-world applications. To address these issues, we propose Adaptive Graph Encoder (AGE), a novel attributed graph embedding framework. AGE consists of two modules: (1) To better alleviate the high-frequency noises in the node features, AGE first applies a carefully-designed Laplacian smoothing filter. (2) AGE employs an adaptive encoder that iteratively strengthens the filtered features for better node embeddings. We conduct experiments using four public benchmark datasets to validate AGE on node clustering and link prediction tasks. Experimental results show that AGE consistently outperforms state-of-the-artgraph embedding methods considerably on these tasks.

Two characteristics of location-based services are mobile trajectories and the ability to facilitate social networking. The recording of trajectory data contributes valuable resources towards understanding users’ geographical movement behaviors. Social networking is possible when users are able to quickly connect to anyone nearby. A social network with location based services is known as location-based social network (LBSN). As shown in Cho et al. [2013], locations that are frequently visited by socially related persons tend to be correlated, which indicates the close association between social connections and trajectory behaviors of users in LBSNs. To better analyze and mine LBSN data, we need to have a comprehensive view of each of these two aspects, i.e., the mobile trajectory data and the social network. Specifically, we present a novel neural network model that can jointly model both social networks and mobile trajectories. Our model consists of two components: the construction of social networks and the generation of mobile trajectories. First we adopt a network embedding method for the construction of social networks: a networking representation can be derived for a user. The key to our model lies in generating mobile trajectories. Second, we consider four factors that influence the generation process of mobile trajectories: user visit preference, influence of friends, short-term sequential contexts, and long-term sequential contexts. To characterize the last two contexts, we employ the RNN and GRU models to capture the sequential relatedness in mobile trajectories at the short or long term levels. Finally, the two components are tied by sharing the user network representations. Experimental results on two important applications demonstrate the effectiveness of our model. In particular, the improvement over baselines is more significant when either network structure or trajectory data is sparse.