Temporal causal discovery aims to uncover causal relations in time series data. Current deep learning-based methods usually analyze the parameters of some components of the trained models, which is an incomplete mapping process from the model parameters to the causality and fails to investigate the other components. To address this, this paper presents an interpretable transformer-based causal discovery model termed CausalFormer, which consists of: 1) the causality-aware transformer which learns the causal representation with the multi-kernel causal convolution under the temporal priority constraint, and 2) the decomposition-based causality detector which identifies causality by interpreting the global structure of the trained transformer with the regression relevance propagation.

Accurate prediction of Sea Surface Temperature (SST) is of high importance in marine science, benefiting applications ranging from ecosystem protection to extreme weather forecasting and climate analysis. Wide-area SST usually shows diverse SST patterns in different sea areas due to the changes of temperature zones and the dynamics of ocean currents. However, existing studies on SST prediction often focus on small-area predictions and lack the consideration of diverse SST patterns. Furthermore, SST shows an annual periodicity, but the periodicity is not strictly adherent to an annual cycle. Existing SST prediction methods struggle to adapt to this non-strict periodicity. To address these two issues, we proposed the Cross-Region Graph Convolutional Network with Periodicity Shift Adaptation (RGCN-PSA) model which is equipped with the Cross-Region Graph Convolutional Network module and the Periodicity Shift Adaption module. The Cross-Region Graph Convolutional Network module enhances wide-area SST prediction by learning and incorporating diverse SST patterns. Meanwhile, the periodicity Shift Adaptation module accounts for the annual periodicity and enable the model to adapt to the possible temporal shift automatically. We conduct experiments on two real-world SST datasets, and the results demonstrate that our RGCN-PSA model obviously outperforms baseline models in terms of prediction accuracy. The code of RGCN-PSA model is available at https://github.com/ADMIS-TONGJI/RGCN-PSA/ .

With the growing diversity of data formats on social media, such as text, images, and videos, there is a growing need to analyze sentiment from multiple modalities. Multimodal sentiment detection, which aims to identify users’ sentiment by jointly modeling information from different modalities, has thus attracted increasing attention. However, most existing multimodal sentiment detection methods fuse multimodal information directly after the unimodal encoding and overlook the modality consistency of multimodal vector spaces before the fusion, which may damage the accuracy of multimodal sentiment detection. To address this issue, we propose a contrastive learning-based multimodal sentiment detection model termed EPMC which can map the representations of different modalities into a unified semantic space before fusion. EPMC operates in two stages, i.e., pre-training stage and fine-tuning stage. At the pre-training stage, we designed a cross-modal transformation module to map different modalities into a unified feature space. Meanwhile, to further capture the relationship between the cross-modal transformation vectors and the unimodal encoding vectors, we propose a multimodal consistency contrastive learning task that helps the model discern and amplify the cross-modal similarity between different modalities, thereby learning more discriminative features for sentiment detection. At the fine-tuning stage, EPMC is iteratively refined using the learned multimodal representation and guided by the cross-entropy loss. Extensive experiments conducted on three public multimodal datasets validate the effectiveness of EPMC model. The official implementation of EPMC is released at https://github.com/ADMIS-TONGJI/EPMC .

Given a specific region, crime prediction aims to predict the occurrence of various crime events within a certain period of time in future, which is of high significance for guaranteeing urban safety. In practice, crime events are usually affected by a variety of factors from different views, e.g., the attributes of the region, the correlations between different regions, and the correlations between different categories of crime events. Moreover, these correlations are dynamically changing over time, which makes it difficult to learn the regularity and patterns in crime data for achieving accurate prediction. To address this issue, we proposed a new M ulti- V iew S patial- T emporal (MVST) model for fine-grained crime prediction. MVST model first builds a static region graph to capture the similarity between regions in terms of region attributes such as census records and economy statistics, and creates a time-dependent graph to capture the dynamic correlations between regions based on human mobility data. Meanwhile, both static and dynamic graphs are created to capture the correlations between different categories of crime events. After that, those graphs created from different views are fused together with a multi-view graph fusion module to achieve crime prediction with fine-grained time granularities, e.g., 4 hours and 12 hours. According to the experiments on two real crime datasets, our MVST model obviously outperforms existing crime prediction methods. The code of MVST model is available at https://github.com/weichang811/MVST .

Inductive relation prediction, an important task for knowledge graph completion, is to predict the relations between entities that are unseen at the training stage. The latest methods use Pre-Trained Language Models (PLMs) to encode the paths between the head entity and tail entity and achieve state-of-the-art prediction performance. However, these methods cannot handle no-path scenarios well and lack the capability to learn comprehensive relation representations for distinguishing different relations. To tackle this issue, we propose a novel R elation- a ware k nowledg e r easoning model entitled Raker, which introduces an adaptive reasoning information extraction method to identify relation-aware reasoning neighbors of entities in the target triple to handle no-path scenarios and enables the PLM to better distinguish different relations via the relation-specific soft prompting. Raker is evaluated on three public datasets and achieves SOTA performance in inductive relation prediction when compared with the baseline methods. Notably, the absolute improvement of Raker is even more than 5% on the FB15k-237 dataset in the inductive setting. Moreover, Raker also demonstrates the superiority in transductive, few-shot, and unseen relation settings. The code of Raker is available at https://github.com/ADMIS-TONGJI/Raker .

Accurate spatial-temporal (ST) traffic prediction plays an essential role in intelligent transportation systems. Existing advanced traffic prediction methods typically utilize spatial-temporal graph neural networks (STGNNs) to capture the ST correlations and achieve excellent prediction performance. However, our experimental investigation reveals that existing static and dynamic graph-based STGNNs still incur excessive noise and redundancy, and fail to discover robust and reliable ST correlations in traffic networks. Moreover, most methods cannot explain the underlying reasons behind the ST correlations. To solve these problems, we propose a novel S patial- T emporal G raph M odeling framework via A daptive contrastive learning (ST-GMA). Firstly, we design a robust augmentation learning module to generate high-level and robust data augmentations via a self-supervised task for modeling reliable correlations. Then, we develop an adaptive contrastive learning module to update correlation graphs by effectively selecting positive and negative augmentations, reducing redundant calculations, and providing insights into the correlation changes. Finally, ST-GMA integrates the generated correlation graphs with ST convolution blocks to conduct traffic prediction tasks. Experimental results on five real-world datasets demonstrate that ST-GMA not only achieves significant prediction performance compared with state-of-the-art methods but also exhibits a new perspective on the interpretability of correlation changes.

Temporal causal discovery is a crucial task aimed at uncovering the causal relations within time series data. The latest temporal causal discovery methods usually train deep learning models on prediction tasks to uncover the causality between time series. They capture causal relations by analyzing the parameters of some components of the trained models, e.g., attention weights and convolution weights. However, this is an incomplete mapping process from the model parameters to the causality and fails to investigate the other components, e.g., fully connected layers and activation functions, that are also significant for causal discovery. To facilitate the utilization of the whole deep learning models in temporal causal discovery, we proposed an interpretable transformer-based causal discovery model termed CausalFormer, which consists of the causality-aware transformer and the decomposition-based causality detector. The causality-aware transformer learns the causal representation of time series data using a prediction task with the designed multi-kernel causal convolution which aggregates each input time series along the temporal dimension under the temporal priority constraint. Then, the decomposition-based causality detector interprets the global structure of the trained causality-aware transformer with the proposed regression relevance propagation to identify potential causal relations and finally construct the causal graph. Experiments on synthetic, simulated, and real datasets demonstrate the state-of-the-art performance of CausalFormer on discovering temporal causality.