Molecule graph editing has become a powerful paradigm for optimizing chemical compounds in drug discovery. Existing methods overlook the invariant structure-property relationships, and rely on variable correlations that shift across different instructions, thereby failing to generalize to out-of-distribution (O.O.D.) scenarios. To overcome the weakness of existing work, in this paper we propose to capture and utilize the invariant factors in order to achieve generalizable molecule graph editing under distribution shifts. However, this problem remains challenging, given that the invariant and variant factors are deeply entangled within the editing models. To tackle this challenge, we propose MoFE, a disentangled graph large language model for molecule graph editing that handles editing instructions under distribution shifts via disentangling invariant factors that govern editing-relevant properties. Specifically, we propose a disentangled graph projector with invariance loss that encodes molecular graphs into disentangled latent factors, with an invariance loss that ensures consistency across paraphrased prompts with the same objective. Then, we enhance the LLM with a factor-aware LoRA mixture-of-experts, where each expert is associated with a distinct latent factor. Additionally, we introduce a factor disentanglement loss weighting strategy that adaptively assigns higher weights to expert-factor pairs that perform well on relevant editing tasks. The proposed MoFE model promotes joint disentanglement between experts and latent factors, reinforcing their alignment and preventing collapse. Experiments on a representative benchmark demonstrate that MoFE is able to achieve superior O.O.D. generalization performance in molecule graph editing.

Dynamic graph neural networks, i.e., DyGNNs, have been extensively explored in literature to handle structural and temporal properties in graphs. On the one hand, there naturally exist distribution shifts in real-world scenarios relevant to dynamic graphs. On the other hand, the dynamics may further bring extra uncertainties to patterns in dynamic graphs. However, existing DyGNNs merely exploit variant patterns with respect to labels under distribution shifts, failing to accurately make predictions when there exist distribution shifts together with uncertain patterns from training data to test data. To deal with this issue, in this paper we propose to handle spatio-temporal distribution shifts in dynamic graphs via the discovery and utilization of invariant patterns, taking uncertainties in patterns into account, where the invariant patterns include structures and features whose predictive abilities are stable across distribution shifts. Nevertheless, we face the following key challenges: i) How to discover the complex invariant and variant spatio-temporal patterns involving time-varying topological structures and node-level features; ii) How to utilize the invariant and variant patterns to deal with the spatio-temporal distribution shifts in dynamic graphs; iii) How to handle the pattern uncertainties upon capturing the hidden invariance and variance with a theoretical guarantee. To tackle these challenges, we propose the Information Bottleneck guided Disentangled Dynamic Graph ATtention network (IB-D$^{2}$2 GAT). Our proposed IB-D$^{2}$2 GAT model is able to effectively handle spatio-temporal distribution shifts with uncertainties in dynamic graphs through discovering variant and invariant spatio-temporal patterns via information bottleneck. Specifically, we propose a disentangled spatio-temporal attention network to capture the invariant and variant patterns. Next, guided by the information bottleneck principle, we propose the distribution-based invariance optimization strategy which injects stochasticity into the invariant pattern identification so as to prevent the variant information from influencing the prediction, thus eliminating the spurious impacts of variant patterns. We further theoretically show that our proposed tailored invariance optimization strategy can lead to accurately capturing the invariant patterns with stable predictive abilities and therefore is capable of handling distribution shifts. Experiments on multiple real-world datasets and one synthetic dataset demonstrate the superiority of our method over state-of-the-art baselines under distribution shifts.

Deep neural networks have achieved great success in the last decade. When designing neural networks to handle the ubiquitous geometric data such as point clouds and graphs, it is critical that the model can maintain invariance towards various transformations such as translation, rotation, and scaling. Most existing graph neural network (GNN) approaches can only maintain permutation-invariance, failing to guarantee invariance with respect to other transformations. Besides GNNs, other works design sophisticated transformation-invariant layers, which are computationally expensive and difficult to be extended. In this paper, we revisit why general neural networks cannot maintain transformation invariance. Our findings show that transformation-invariant and distance-preserving initial point representations are sufficient to achieve transformation invariance rather than needing sophisticated neural layer designs. Motivated by these findings, we propose Transformation Invariant Neural Networks (TinvNet), a straightforward and general plug-in for geometric data. Specifically, we realize transformation invariant and distance-preserving initial point representations by modifying multi-dimensional scaling and feed the representations into existing neural networks. We prove that TinvNet can strictly guarantee transformation invariance, being general and flexible enough to be combined with the existing neural networks. Extensive experimental results on point cloud analysis and combinatorial optimization demonstrate the effectiveness and general applicability of our method. We also extend our method into equivariance cases. Based on the results, we advocate that TinvNet should be considered as an essential baseline for further studies of transformation-invariant geometric deep learning.

Video generation has witnessed great success recently, but its application in generating long videos still remains challenging due to the difficulty in maintaining the temporal consistency of generated videos and the high memory cost during generation. To tackle the problems, in this paper, we propose a brave and new idea of Multi-sentence Video Grounding for Long Video Generation, connecting the massive video moment retrieval to the video generation task for the first time, providing a new paradigm for long video generation. The method of our work can be summarized as three steps: (i) We design sequential scene text prompts as the queries for video grounding, utilizing the massive video moment retrieval to search for video moment segments that meet the text requirements in the video database. (ii) Based on the source frames of retrieved video moment segments, we adopt video editing methods to create new video content while preserving the temporal consistency of the retrieved video. Since editing can be conducted segment by segment and even frame by frame, it largely reduces memory costs. (iii) We also attempt video morphing and personalized generation methods to improve the subject consistency of long video generation, providing ablation experimental results for the subtasks of long video generation. Our approach seamlessly extends the development in image/video editing, video morphing and personalized generation, and video grounding to the long video generation, offering effective solutions for generating long videos at a low memory cost.

Self-supervised learning (SSL) on graph-structured data has attracted considerable attention recently. Masked graph autoencoder, as one promising generative graph SSL approach that aims to recover masked parts of the input graph data, has shown great success in various downstream graph tasks. However, existing masked graph autoencoders fail to consider the degree of difficulty of recovering the masked edges that often have different impacts on the model performance, resulting in suboptimal node representations. To tackle this challenge, in this paper, we propose a novel curriculum based self-supervised masked graph autoencoder that is able to capture and leverage the underlying degree of difficulty of data dependencies hidden in edges, and design better mask-reconstruction pretext tasks for learning informative node representations. Specifically, we first design a difficulty measurer to identify the underlying structural degree of difficulty of edges during the masking step. Then, we adopt a self-paced scheduler to determine the order of masking edges, which encourages the graph encoder to learn from easy to difficult parts. Finally, the masked edges are gradually incorporated into the reconstruction pretext task, leading to high-quality node representations. Experiments on several real-world node classification and link prediction datasets demonstrate the superiority of our proposed method over state-of-the-art graph self-supervised learning baselines. This work is the first study of curriculum strategy for masked graph autoencoders, to the best of our knowledge.

Graph foundation models (GFMs) aim to share graph knowledge across diverse domains and tasks to boost graph machine learning. However, existing GFMs rely on hand-designed and fixed graph neural network (GNN) architectures, failing to utilize optimal architectures *w.r.t.* specific domains and tasks, inevitably leading to suboptimal performance in diverse graph domains and tasks. In this paper, we explore graph neural architecture search (GNAS) for GFMs for the first time, which suffers from the problem of *architecture inconsistency*, i.e., the optimal architectures for different tasks and domains vary. We tackle this problem by discovering an invariant graph-architecture relationship across domains and tasks, which imposes three challenges: i) how to capture invariant and variant patterns; ii) how to customize architectures to adapt to diverse domains and tasks; iii) how to mitigate the data domination phenomenon during the architecture search process. To address these challenges, we propose **Auto**mated **G**raph **F**oundation **M**odel with Adaptive Architecture Customization (**AutoGFM**), providing a theoretical analysis to demonstrate the limitations of existing GNAS. Specifically, we first propose a disentangled contrastive graph encoder to learn invariant and variant patterns. Then, we design an invariant-guided architecture customization strategy to customize architectures for data from diverse domains and tasks. Finally, we propose a curriculum architecture customization mechanism to mitigate the phenomenon of particular data dominating the search process. Extensive experiments demonstrate that **AutoGFM** outperforms baselines, achieving state-of-the-art performance.

Continual multimodal instruction tuning is crucial for adapting Multimodal Large Language Models (MLLMs) to evolving tasks. However, most existing methods adopt a fixed architecture, struggling with adapting to new tasks due to static model capacity. We propose to evolve the architecture under parameter budgets for dynamic task adaptation, which remains unexplored and imposes two challenges: 1) task architecture conflict, where different tasks require varying layer-wise adaptations, and 2) modality imbalance, where different tasks rely unevenly on modalities, leading to unbalanced updates. To address these challenges, we propose a novel Dynamic Mixture of Curriculum LoRA Experts (D-MoLE) method, which automatically evolves MLLM's architecture with controlled parameter budgets to continually adapt to new tasks while retaining previously learned knowledge. Specifically, we propose a dynamic layer-wise expert allocator, which automatically allocates LoRA experts across layers to resolve architecture conflicts, and routes instructions layer-wisely to facilitate knowledge sharing among experts. Then, we propose a gradient-based inter-modal continual curriculum, which adjusts the update ratio of each module in MLLM based on the difficulty of each modality within the task to alleviate the modality imbalance problem. Extensive experiments show that D-MoLE significantly outperforms state-of-the-art baselines, achieving a 15 percent average improvement over the best baseline. To the best of our knowledge, this is the first study of continual learning for MLLMs from an architectural perspective.

Graph neural architecture search (NAS) has emerged as a promising approach for autonomously designing graph neural network architectures by leveraging correlations between graphs and architectures. However, existing methods merely rely on correlations, which may be spurious and vary across distributions. This reliance, without considering causal graph-architecture relationships, limits their ability to generalize under distribution shifts that are ubiquitous in real-world graph scenarios. In this paper, we propose to handle the distribution shifts in NAS process by exploiting the causal graph-architecture relationship to search for optimal architectures that can generalize under distribution shifts. Key challenges remain unexplored: discovering causal graph-architecture relationships with stable cross-distribution predictive abilities, and leveraging them to handle distribution shifts. To address these challenges, we propose a novel approach, Causal-aware Graph Neural Architecture Search (CARNAS), which is capable of capturing causal graph-architecture relationship during NAS process and discovering optimal graph architecture under distribution shifts. We propose Disentangled Causal Subgraph Identification to extract causal subgraphs with stable predictive power across distributions, followed by Graph Embedding Intervention to intervene on these subgraphs in latent space by preserving essential features while filtering out non-causal elements, and Invariant Architecture Customization to enhance their causal invariance for optimizing graph architectures. Extensive experiments on synthetic and real-world datasets show that CARNAS enhances out-of-distribution generalization by uncovering causal graph-architecture relationships during NAS.

Drug-drug interaction (DDI) prediction is a pivotal task in biomedical research. Emerging multimodal approaches that integrate graph neural networks (GNNs) and large language models (LLMs) have gained traction, as GNNs capture molecular structures while LLMs provide a rich biomedical context. However, real-world DDI data often exhibit distribution shifts across structural and textual dimensions, stemming from variations in molecular scaffolds, drug sizes, and assay conditions. Existing methods assume an independent and identically distributed (I.I.D.) setting, failing to handle such shifts primarily due to there key limitations: (i) the entanglement of core interaction motifs with incidental structural features; (ii) inflexible message-passing GNN architectures ill-suited for diverse drug pairs; and (iii) underutilized biomedical knowledge in LLMs for capturing pairwise interaction semantics. These limitations highlight the need for a disentangled, dynamic, and pairwise-aware modeling strategy to achieve out-of-distribution generalized DDI prediction. To solve this problem, we propose DyNamic Pairwise Architecture Search for Generalizable Drug-Drug Interaction LLM (DyNAS-DDI), a novel framework that dynamically adapts network architectures for each molecular pair and integrates biomedical knowledge from LLMs to improve generalization under distribution shifts. Specifically, we propose three modules: (i) Motif-driven disentangled molecule encoding, which disentangles molecular representations into distinct motif-based features while preserving key structural signals through a self-supervised graph encoder; (ii) Attentionbased pairwise neural architecture search, where multi-head attention enriches molecular features to guide a dynamic search mechanism that adaptively optimizes message passing for diverse interaction types; and (iii) retrieval-augmented molecular instruction tuning, where external biomedical knowledge is incorporated to improve interpretability and enable reasoning for unseen drug interactions. Extensive experiments on four datasets for DDI with out-of-distribution (OOD) splits demonstrate our method's superior generalization abilities under distribution shifts. Our code can be available at https://github.com/EkkoXiao/DyNAS-DDI.

Multi-conditional image generation aims to create customized images that align with multiple specified conditions. Existing methods, whether through end-to-end training or by fine-tuning adapters to integrate pre-trained control modules of the same category (e.g., LoRA, IP-Adapter, ControlNet, T2I-Adapter), are restricted to a closed set of predefined input conditions. To overcome this limitation, we propose ModuleTeam, a training-free method for latent mixture of arbitrary control modules, capable of handling open-set conditions by incorporating the corresponding modules. The design of ModuleTeam is rooted in two key findings: (i) modules interfere with each other at the level of model parameters, and (ii) module weights contribute to the generated images by affecting the noise predictions within the diffusion process in an approximately linear manner. The first finding motivates our latent mixture approach, which mixes the control modules by aggregating their latent variables between diffusion model blocks. The second finding enables a multi-inference module reweighting strategy that balances module contributions to generation, requiring no additional training or fine-tuning overhead. Extensive results demonstrate that ModuleTeam not only outperforms existing methods but also provides flexibility in the types of conditions and scalability in their number.

Graph anomaly detection (GAD) is widely prevalent in scenarios such as financial fraud detection, anti-money laundering, and social bot detection. However, structural distribution shifts are commonly observed in real-world GAD data due to selection bias, resulting in reduced homophily. Existing GAD methods tend to rely on homophilic shortcuts when trained on high-homophily structures, limiting their ability to generalize well to data with low homophily under structural distribution shifts. In this study, we propose to handle structural distribution shifts by generating novel environments characterized by diverse homophilic structures and utilizing invariant patterns, i.e., features and structures with the capability of stable prediction across structural distribution shifts, which face two challenges: (1) How to discover invariant patterns from entangled features and structures, as structures are sensitive to varying homophilic distributions. (2) How to systematically construct new environments with diverse homophilic structures. To address these challenges, we propose the Ego-Neighborhood Disentangled Encoder with Homophily-aware Environment Mixup (HEM), which effectively handles structural distribution shifts in GAD by discovering invariant patterns. Specifically, we first propose an ego-neighborhood disentangled encoder to decouple the learning of feature embeddings and structural embeddings, which facilitates subsequent improvements in the invariance of structural embeddings for prediction. Next, we introduce a homophily-aware environment mixup that dynamically adjusts edge weights through adversarial learning, effectively generating environments with diverse structural distributions. Finally, we iteratively train the classifier and environment mixup via adversarial training, simultaneously improving the diversity of constructed environments and discovering invariant patterns under structural distribution shifts. Extensive experiments on real-world datasets demonstrate that our method outperforms existing baselines and achieves state-of-the-art performance under structural distribution shift conditions.

Sequential recommendation, leveraging user-item interaction histories to provide personalized and timely suggestions, has drawn significant research interest recently. With the power of exploiting spatio-temporal dynamics, Dynamic Graph Neural Networks (DyGNNs) show great potential in sequential recommendation by modeling the dynamic relationship between users and items. However, spatio-temporal distribution shifts naturally exist in out-of-distribution sequential recommendation, where both user-item relationships and temporal sequences demonstrate pattern shifts. The out-of-distribution scenarios may lead to the failure of existing DyGNNs in handling spatio-temporal distribution shifts in sequential recommendation, given that the patterns they exploit tend to be variant w.r.t labels under distribution shifts. In this article, we propose Disentangled Intervention-based Dynamic graph Attention networks with Invariance Promotion ( I-DIDA ) to handle spatio-temporal distribution shifts in sequential recommendation by discovering and utilizing invariant patterns , i.e., structures and features whose predictive abilities are stable across distribution shifts. Specifically, we first propose a disentangled spatio-temporal attention network to capture the variant and invariant patterns. By utilizing the disentangled patterns, we design a spatio-temporal intervention mechanism to create multiple interventional distributions and an environment inference module to infer the latent spatio-temporal environments, and minimize the invariance loss to leverage the invariant patterns with stable predictive abilities under distribution shifts. Extensive experiments demonstrate the superiority of our method over state-of-the-art sequential recommendation baselines under distribution shifts.

In the field of Artificial Intelligence for Information Technology Operations, causal discovery is pivotal for operation and maintenance of systems, facilitating downstream industrial tasks such as root cause analysis. Temporal causal discovery, as an emerging method, aims to identify temporal causal relations between variables directly from observations by utilizing interventional data. However, existing methods mainly focus on synthetic datasets with heavy reliance on interventional targets and ignore the textual information hidden in real-world systems, failing to conduct causal discovery for real industrial scenarios. To tackle this problem, in this paper we investigate temporal causal discovery in industrial scenarios, which faces two critical challenges: how to discover causal relations without the interventional targets that are costly to obtain in practice, and how to discover causal relations via leveraging the textual information in systems which can be complex yet abundant in industrial contexts. To address these challenges, we propose the RealTCD framework, which is able to leverage domain knowledge to discover temporal causal relations without interventional targets. We first develop a score-based temporal causal discovery method capable of discovering causal relations without relying on interventional targets through strategic masking and regularization. Then, by employing Large Language Models (LLMs) to handle texts and integrate domain knowledge, we introduce LLM-guided meta-initialization to extract the meta-knowledge from textual information hidden in systems to boost the quality of discovery. We conduct extensive experiments on both simulation datasets and our real-world application scenario to show the superiority of our proposed RealTCD over existing baselines in temporal causal discovery.

Graph out-of-distribution (OOD) generalization, aiming to generalize graph neural networks (GNNs) under distribution shifts between training and testing environments, has attracted ever-increasing attention recently. However, existing literature heavily relies on sufficient task-dependent graph labels, which are often scarce or even unavailable, limiting their applications in real-world scenarios. In this paper, we study the self-supervised graph OOD generalization problem, i.e., learning GNNs capable of achieving relatively stable performances under distribution shifts without graph labels. However, the problem remains largely unexplored, with the critical challenge that the invariant and variant information are highly entangled in graphs. To solve this problem, we propose an OOD generalized disentangled graph contrastive learning model (OOD-GCL), which is capable of learning disentangled graph-level representations with self-supervision that can handle distribution shifts between training and testing graph data. Specifically, we first introduce a disentangled graph encoder to map each input graph into the factorized graph representation. Then we propose a tailored disentangled invariant self-supervised learning module to maximize predictive ability of the representations and make sure the representations other than from one specific channel are invariant to the environments partitioned by this latent factor for excluding the information corresponding to this latent factor for disentanglement. Finally, the disentangled graph representations are fed into a linear predictor and finetuned for the downstream tasks. We provide comprehensive theoretical analyses to show that our model can learn disentangled graph representations and achieve OOD generalization. Extensive experiments on real-world datasets demonstrate the superiority of our model against state-of-the-art baselines under distribution shifts for graph classification tasks.

The existing graph neural architecture search (GNAS) methods assume that the graph tasks are static during the search process, ignoring the ubiquitous scenarios where sequential graph tasks come in a continual fashion. Moreover, existing GNAS works resort to entangled graph factors during the architecture search process, resulting in the catastrophic forgetting problems. In this paper, we study the problem of continual graph neural architecture search that is expected to continually search the architecture to learn new graph tasks without forgetting the past, which remains largely unexplored in the literature. However, this problem poses the challenge of architecture conflicts, i.e., the optimal architecture for the new graph task may have performance deterioration and thus sub-optimal for past tasks. To address the challenge, we propose a novel Disentangled Continual Graph Neural Architecture Search with Invariant Modularization (GASIM) method, which is able to continually search the optimal architectures without forgetting past knowledge. Specifically, we first design a modular graph architecture super-network incorporating multiple modules to enable searching architecture with factor expertise. Second, we propose a factor-based task-module router that discovers the latent graph factors and routes the incoming task to the best suitable architecture module to alleviate the forgetting problem induced by architecture conflicts. Finally, we propose an invariant architecture search mechanism to capture the shared knowledge among tasks. Extensive experiments on real-world datasets demonstrate that the proposed method achieves state-of-the-art performance against baselines in continual graph neural architecture search.

Graph Neural Architecture Search (GNAS) has achieved superior performance on various graph-structured tasks. However, existing GNAS studies overlook the applications of GNAS in resource-constraint scenarios. This paper proposes to design a joint graph data and architecture mechanism, which identifies important sub-architectures via the valuable graph data. To search for optimal lightweight Graph Neural Networks (GNNs), we propose a Lightweight Graph Neural Architecture Search with Graph SparsIfication and Network Pruning (GASSIP) method. In particular, GASSIP comprises an operation-pruned architecture search module to enable efficient lightweight GNN search. Meanwhile, we design a novel curriculum graph data sparsification module with an architecture-aware edge-removing difficulty measurement to help select optimal sub-architectures. With the aid of two differentiable masks, we iteratively optimize these two modules to efficiently search for the optimal lightweight architecture. Extensive experiments on five benchmarks demonstrate the effectiveness of GASSIP. Particularly, our method achieves on-par or even higher node classification performance with half or fewer model parameters of searched GNNs and a sparser graph.

In an era marked by the increasing adoption of Large Language Models (LLMs) for various tasks, there is a growing focus on exploring LLMs' capabilities in handling web data, particularly graph data. Dynamic graphs, which capture temporal network evolution patterns, are ubiquitous in real-world web data. Evaluating LLMs' competence in understanding spatial-temporal information on dynamic graphs is essential for their adoption in web applications, which remains unexplored in the literature. In this paper, we bridge the gap via proposing to evaluate LLMs' spatial-temporal understanding abilities on dynamic graphs, to the best of our knowledge, for the first time. Specifically, we propose the LLM4DyG benchmark, which includes nine specially designed tasks considering the capability evaluation of LLMs from both temporal and spatial dimensions. Then, we conduct extensive experiments to analyze the impacts of different data generators, data statistics, prompting techniques, and LLMs on the model performance. Finally, we propose Disentangled Spatial-Temporal Thoughts (DST2) for LLMs on dynamic graphs to enhance LLMs' spatial-temporal understanding abilities. Our main observations are: 1) LLMs have preliminary spatial-temporal understanding abilities on dynamic graphs, 2) Dynamic graph tasks show increasing difficulties for LLMs as the graph size and density increase, while not sensitive to the time span and data generation mechanism, 3) the proposed DST2 prompting method can help to improve LLMs' spatial-temporal understanding abilities on dynamic graphs for most tasks. The data and codes are publicly available at Github.

Generating customized content in videos has received increasing attention recently. However, existing works primarily focus on customized text-to-video generation for single subject, suffering from subject-missing and attribute-binding problems when the video is expected to contain multiple subjects. Furthermore, existing models struggle to assign the desired actions to the corresponding subjects (action-binding problem), failing to achieve satisfactory multi-subject generation performance. To tackle the problems, in this paper, we propose DisenStudio, a novel framework that can generate text-guided videos for customized multiple subjects, given few images for each subject. Specifically, DisenStudio enhances a pretrained diffusion-based text-to-video model with our proposed spatial-disentangled cross-attention mechanism to associate each subject with the desired action. Then the model is customized for the multiple subjects with the proposed motion-preserved disentangled finetuning, which involves three tuning strategies: multi-subject co-occurrence tuning, masked single-subject tuning, and multi-subject motion-preserved tuning. The first two strategies guarantee the subject occurrence and preserve their visual attributes, and the third strategy helps the model maintain the temporal motion-generation ability when finetuning on static images. We conduct extensive experiments to demonstrate our proposed DisenStudio significantly outperforms existing methods in various metrics. Additionally, we show that DisenStudio can be used as a powerful tool for various controllable generation applications.

Existing Video Corpus Moment Retrieval (VCMR) is limited to coarse-grained understanding that hinders precise video moment localization when given fine-grained queries. In this paper, we propose a more challenging fine-grained VCMR benchmark requiring methods to localize the best-matched moment from the corpus with other partially matched candidates. To improve the dataset construction efficiency and guarantee high-quality data annotations, we propose VERIFIED, an automatic \underline{V}id\underline{E}o-text annotation pipeline to generate captions with \underline{R}el\underline{I}able \underline{FI}n\underline{E}-grained statics and \underline{D}ynamics. Specifically, we resort to large language models (LLM) and large multimodal models (LMM) with our proposed Statics and Dynamics Enhanced Captioning modules to generate diverse fine-grained captions for each video. To filter out the inaccurate annotations caused by the LLM hallucination, we propose a Fine-Granularity Aware Noise Evaluator where we fine-tune a video foundation model with disturbed hard-negatives augmented contrastive and matching losses. With VERIFIED, we construct a more challenging fine-grained VCMR benchmark containing Charades-FIG, DiDeMo-FIG, and ActivityNet-FIG which demonstrate a high level of annotation quality. We evaluate several state-of-the-art VCMR models on the proposed dataset, revealing that there is still significant scope for fine-grained video understanding in VCMR.

Dynamic heterogeneous graph neural networks (DHGNNs) have been shown to be effective in handling the ubiquitous dynamic heterogeneous graphs. However, the existing DHGNNs are hand-designed, requiring extensive human efforts and failing to adapt to diverse dynamic heterogeneous graph scenarios. In this paper, we propose to automate the design of DHGNN, which faces two major challenges: 1) how to design the search space to jointly consider the spatial-temporal dependencies and heterogeneous interactions in graphs; 2) how to design an efficient search algorithm in the potentially large and complex search space. To tackle these challenges, we propose a novel Dynamic Heterogeneous Graph Attention Search (DHGAS) method. Our proposed method can automatically discover the optimal DHGNN architecture and adapt to various dynamic heterogeneous graph scenarios without human guidance. In particular, we first propose a unified dynamic heterogeneous graph attention (DHGA) framework, which enables each node to jointly attend its heterogeneous and dynamic neighbors. Based on the framework, we design a localization space to determine where the attention should be applied and a parameterization space to determine how the attention should be parameterized. Lastly, we design a multi-stage differentiable search algorithm to efficiently explore the search space. Extensive experiments on real-world dynamic heterogeneous graph datasets demonstrate that our proposed method significantly outperforms state-of-the-art baselines for tasks including link prediction, node classification and node regression. To the best of our knowledge, DHGAS is the first dynamic heterogeneous graph neural architecture search method.

Dynamic graph neural networks (DyGNNs) currently struggle with handling distribution shifts that are inherent in dynamic graphs. Existing work on DyGNNs with out-of-distribution settings only focuses on the time domain, failing to handle cases involving distribution shifts in the spectral domain. In this paper, we discover that there exist cases with distribution shifts unobservable in the time domain while observable in the spectral domain, and propose to study distribution shifts on dynamic graphs in the spectral domain for the first time. However, this investigation poses two key challenges: i) it is non-trivial to capture different graph patterns that are driven by various frequency components entangled in the spectral domain; and ii) it remains unclear how to handle distribution shifts with the discovered spectral patterns. To address these challenges, we propose Spectral Invariant Learning for Dynamic Graphs under Distribution Shifts (SILD), which can handle distribution shifts on dynamic graphs by capturing and utilizing invariant and variant spectral patterns. Specifically, we first design a DyGNN with Fourier transform to obtain the ego-graph trajectory spectrums, allowing the mixed dynamic graph patterns to be transformed into separate frequency components. We then develop a disentangled spectrum mask to filter graph dynamics from various frequency components and discover the invariant and variant spectral patterns. Finally, we propose invariant spectral filtering, which encourages the model to rely on invariant patterns for generalization under distribution shifts. Experimental results on synthetic and real-world dynamic graph datasets demonstrate the superiority of our method for both node classification and link prediction tasks under distribution shifts.

The existing graph neural architecture search (GNAS) methods heavily rely on supervised labels during the search process, failing to handle ubiquitous scenarios where supervisions are not available. In this paper, we study the problem of unsupervised graph neural architecture search, which remains unexplored in the literature. The key problem is to discover the latent graph factors that drive the formation of graph data as well as the underlying relations between the factors and the optimal neural architectures. Handling this problem is challenging given that the latent graph factors together with architectures are highly entangled due to the nature of the graph and the complexity of the neural architecture search process. To address the challenge, we propose a novel Disentangled Self-supervised Graph Neural Architecture Search (DSGAS) model, which is able to discover the optimal architectures capturing various latent graph factors in a self-supervised fashion based on unlabeled graph data. Specifically, we first design a disentangled graph super-network capable of incorporating multiple architectures with factor-wise disentanglement, which are optimized simultaneously. Then, we estimate the performance of architectures under different factors by our proposed self-supervised training with joint architecture-graph disentanglement. Finally, we propose a contrastive search with architecture augmentations to discover architectures with factor-specific expertise. Extensive experiments on 11 real-world datasets demonstrate that the proposed model is able to achieve state-of-the-art performance against several baseline methods in an unsupervised manner.

Various neural network models have been proposed to tackle combinatorial optimization problems such as the travelling salesman problem (TSP). Existing learning-based TSP methods adopt a simple setting that the training and testing data are independent and identically distributed. However, the existing literature fails to solve TSP instances when training and testing data have different distributions. Concretely, we find that different training and testing distribution will result in more difficult TSP instances, i.e., the solution obtained by the model has a large gap from the optimal solution. To tackle this problem, in this work, we study learning-based TSP methods when training and testing data have different distributions using adaptive-hardness, i.e., how difficult a TSP instance can be for a solver. This problem is challenging because it is non-trivial to (1) define hardness measurement quantitatively; (2) efficiently and continuously generate sufficiently hard TSP instances upon model training; (3) fully utilize instances with different levels of hardness to learn a more powerful TSP solver. To solve these challenges, we first propose a principled hardness measurement to quantify the hardness of TSP instances. Then, we propose a hardness-adaptive generator to generate instances with different hardness. We further propose a curriculum learner fully utilizing these instances to train the TSP solver. Experiments show that our hardness-adaptive generator can generate instances ten times harder than the existing methods, and our proposed method achieves significant improvement over state-of-the-art models in terms of the optimality gap. The codes are publicly available.

Graph neural architecture search (GraphNAS) has recently aroused considerable attention in both academia and industry. However, two key challenges seriously hinder the further research of GraphNAS. First, since there is no consensus for the experimental setting, the empirical results in different research papers are often not comparable and even not reproducible, leading to unfair comparisons. Secondly, GraphNAS often needs extensive computations, which makes it highly inefficient and inaccessible to researchers without access to large-scale computation. To solve these challenges, we propose NAS-Bench-Graph, a tailored benchmark that supports unified, reproducible, and efficient evaluations for GraphNAS. Specifically, we construct a unified, expressive yet compact search space, covering 26,206 unique graph neural network (GNN) architectures and propose a principled evaluation protocol. To avoid unnecessary repetitive training, we have trained and evaluated all of these architectures on nine representative graph datasets, recording detailed metrics including train, validation, and test performance in each epoch, the latency, the number of parameters, etc. Based on our proposed benchmark, the performance of GNN architectures can be directly obtained by a look-up table without any further computation, which enables fair, fully reproducible, and efficient comparisons. To demonstrate its usage, we make in-depth analyses of our proposed NAS-Bench-Graph, revealing several interesting findings for GraphNAS. We also showcase how the benchmark can be easily compatible with GraphNAS open libraries such as AutoGL and NNI. To the best of our knowledge, our work is the first benchmark for graph neural architecture search.

Dynamic graph neural networks (DyGNNs) have demonstrated powerful predictive abilities by exploiting graph structural and temporal dynamics. However, the existing DyGNNs fail to handle distribution shifts, which naturally exist in dynamic graphs, mainly because the patterns exploited by DyGNNs may be variant with respect to labels under distribution shifts. In this paper, we propose to handle spatio-temporal distribution shifts in dynamic graphs by discovering and utilizing {\it invariant patterns}, i.e., structures and features whose predictive abilities are stable across distribution shifts, which faces two key challenges: 1) How to discover the complex variant and invariant spatio-temporal patterns in dynamic graphs, which involve both time-varying graph structures and node features. 2) How to handle spatio-temporal distribution shifts with the discovered variant and invariant patterns. To tackle these challenges, we propose the Disentangled Intervention-based Dynamic graph Attention networks (DIDA). Our proposed method can effectively handle spatio-temporal distribution shifts in dynamic graphs by discovering and fully utilizing invariant spatio-temporal patterns. Specifically, we first propose a disentangled spatio-temporal attention network to capture the variant and invariant patterns. Then, we design a spatio-temporal intervention mechanism to create multiple interventional distributions by sampling and reassembling variant patterns across neighborhoods and time stamps to eliminate the spurious impacts of variant patterns. Lastly, we propose an invariance regularization term to minimize the variance of predictions in intervened distributions so that our model can make predictions based on invariant patterns with stable predictive abilities and therefore handle distribution shifts. Experiments on three real-world datasets and one synthetic dataset demonstrate the superiority of our method over state-of-the-art baselines under distribution shifts. Our work is the first study of spatio-temporal distribution shifts in dynamic graphs, to the best of our knowledge.

Discovering ideal Graph Neural Networks (GNNs) architectures for different tasks is labor intensive and time consuming. To save human efforts, Neural Architecture Search (NAS) recently has been used to automatically discover adequate GNN architectures for certain tasks in order to achieve competitive or even better performance compared with manually designed architectures. However, existing works utilizing NAS to search GNN structures fail to answer the question: how NAS is able to select the desired GNN architectures? In this paper, we investigate this question to solve the problem, for the first time. We conduct a measurement study with experiments to discover that gradient based NAS methods tend to select proper architectures based on the usefulness of different types of information with respect to the target task. Our explorations further show that gradient based NAS also suffers from noises hidden in the graph, resulting in searching suboptimal GNN architectures. Based on our findings, we propose a Graph differentiable Architecture Search model with Structure Optimization (GASSO), which allows differentiable search of the architecture with gradient descent and is able to discover graph neural architectures with better performance through employing graph structure learning as a denoising process in the search procedure. The proposed GASSO model is capable of simultaneously searching the optimal architecture and adaptively adjusting graph structure by jointly optimizing graph architecture search and graph structure denoising. Extensive experiments on real-world graph datasets demonstrate that our proposed GASSO model is able to achieve state-of-the-art performance compared with existing baselines.