Mental-health risk detection seeks early signs of distress from social media posts and clinical transcripts to enable timely intervention before crises.When such risks go undetected,consequences can escalate to self-h...Mental-health risk detection seeks early signs of distress from social media posts and clinical transcripts to enable timely intervention before crises.When such risks go undetected,consequences can escalate to self-harm,long-term disability,reduced productivity,and significant societal and economic burden.Despite recent advances,detecting risk from online text remains challenging due to heterogeneous language,evolving semantics,and the sequential emergence of new datasets.Effective solutions must encode clinically meaningful cues,reason about causal relations,and adapt to new domains without forgetting prior knowledge.To address these challenges,this paper presents a Continual Neuro-Symbolic Graph Learning(CNSGL)framework that unifies symbolic reasoning,causal inference,and continual learning within a single architecture.Each post is represented as a symbolic graph linking clinically relevant tags to textual content,enriched with causal edges derived from directional Point-wise Mutual Information(PMI).A two-layer Graph Convolutional Network(GCN)encodes these graphs,and a Transformer-based attention pooler aggregates node embeddings while providing interpretable tag-level importances.Continual adaptation across datasets is achieved through the Multi-Head Freeze(MH-Freeze)strategy,which freezes a shared encoder and incrementally trains lightweight task-specific heads(small classifiers attached to the shared embedding).Experimental evaluations across six diverse mental-health datasets ranging from Reddit discourse to clinical interviews,demonstrate that MH-Freeze consistently outperforms existing continual-learning baselines in both discriminative accuracy and calibration reliability.Across six datasets,MH-Freeze achieves up to 0.925 accuracy and 0.923 F1-Score,with AUPRC≥0.934 and AUROC≥0.942,consistently surpassing all continual-learning baselines.The results confirm the framework’s ability to preserve prior knowledge,adapt to domain shifts,and maintain causal interpretability,establishing CNSGL as a promising step toward robust,explainable,and lifelong mental-health risk assessment.展开更多
This study proposes a novel forecasting framework that simultaneously captures the strong periodicity and irregular meteorological fluctuations inherent in solar radiation time series.Existing approaches typically def...This study proposes a novel forecasting framework that simultaneously captures the strong periodicity and irregular meteorological fluctuations inherent in solar radiation time series.Existing approaches typically define inter-regional correlations using either simple correlation coefficients or distance-based measures when applying spatio-temporal graph neural networks(STGNNs).However,such definitions are prone to generating spurious correlations due to the dominance of periodic structures.To address this limitation,we adopt the Elastic-Band Transform(EBT)to decompose solar radiation into periodic and amplitude-modulated components,which are then modeled independently with separate graph neural networks.The periodic component,characterized by strong nationwide correlations,is learned with a relatively simple architecture,whereas the amplitude-modulated component is modeled with more complex STGNNs that capture climatological similarities between regions.The predictions from the two components are subsequently recombined to yield final forecasts that integrate both periodic patterns and aperiodic variability.The proposed framework is validated with multiple STGNN architectures,and experimental results demonstrate improved predictive accuracy and interpretability compared to conventional methods.展开更多
Dear Editor,This letter proposes the graph tensor alliance attention network(GT-A^(2)T)to represent a dynamic graph(DG)precisely.Its main idea includes 1)Establishing a unified spatio-temporal message propagation fram...Dear Editor,This letter proposes the graph tensor alliance attention network(GT-A^(2)T)to represent a dynamic graph(DG)precisely.Its main idea includes 1)Establishing a unified spatio-temporal message propagation framework on a DG via the tensor product for capturing the complex cohesive spatio-temporal interdependencies precisely and 2)Acquiring the alliance attention scores by node features and favorable high-order structural correlations.展开更多
Knowledge graphs(KGs),which organize real-world knowledge in triples,often suffer from issues of incompleteness.To address this,multi-hop knowledge graph reasoning(KGR)methods have been proposed for interpretable know...Knowledge graphs(KGs),which organize real-world knowledge in triples,often suffer from issues of incompleteness.To address this,multi-hop knowledge graph reasoning(KGR)methods have been proposed for interpretable knowledge graph completion.The primary approaches to KGR can be broadly classified into two categories:reinforcement learning(RL)-based methods and sequence-to-sequence(seq2seq)-based methods.While each method has its own distinct advantages,they also come with inherent limitations.To leverage the strengths of each method while addressing their weaknesses,we propose a cyclical training method that alternates for several loops between the seq2seq training phase and the policy-based RL training phase using a transformer architecture.Additionally,a multimodal data encoding(MDE)module is introduced to improve the representation of entities and relations in KGs.TheMDE module treats entities and relations as distinct modalities,processing each with a dedicated network specialized for its respective modality.It then combines the representations of entities and relations in a dynamic and fine-grained manner using a gating mechanism.The experimental results from the knowledge graph completion task highlight the effectiveness of the proposed framework.Across five benchmark datasets,our framework achieves an average improvement of 1.7%in the Hits@1 metric and a 0.8%average increase in the Mean Reciprocal Rank(MRR)compared to other strong baseline methods.Notably,the maximum improvement in Hits@1 exceeds 4%,further demonstrating the effectiveness of the proposed approach.展开更多
Computational approaches for predicting drug-target interactions(DTIs)are pivotal in advancing drug discovery.Current methodologies leveraging heterogeneous networks often fall short in fully integrating both local an...Computational approaches for predicting drug-target interactions(DTIs)are pivotal in advancing drug discovery.Current methodologies leveraging heterogeneous networks often fall short in fully integrating both local and global network information.To comprehensively consider network information,we propose DHGT-DTI,a novel deep learning-based approach for DTI prediction.Specifically,we capture the local and global structural information of the network from both neighborhood and meta-path per-spectives.In the neighborhood perspective,we employ a heterogeneous graph neural network(HGNN),which extends Graph Sample and Aggregate(GraphSAGE)to handle diverse node and edge types,effectively learning local network structures.In the meta-path perspective,we introduce a Graph Transformer with residual connections to model higher-order relationships defined by meta-paths,such as"drug-disease-drug",and use an attention mechanism to fuse information across multiple meta-paths.The learned features from these dual perspectives are synergistically integrated for DTI prediction via a matrix decomposition method.Furthermore,DHGT-DTI reconstructs not only the DTI network but also auxiliary networks to bolster prediction accuracy.Comprehensive experiments on two benchmark datasets validate the superiority of DHGT-DTI over existing baseline methods.Additionally,case studies on six drugs used to treat Parkinson's disease not only validate the practical utility of DHGT-DTI but also highlight its broader potential in accelerating drug discovery for other diseases.展开更多
Owing to the constraints of depth sensing technology,images acquired by depth cameras are inevitably mixed with various noises.For depth maps presented in gray values,this research proposes a novel denoising model,ter...Owing to the constraints of depth sensing technology,images acquired by depth cameras are inevitably mixed with various noises.For depth maps presented in gray values,this research proposes a novel denoising model,termed graph-based transform(GBT)and dual graph Laplacian regularization(DGLR)(DGLR-GBT).This model specifically aims to remove Gaussian white noise by capitalizing on the nonlocal self-similarity(NSS)and the piecewise smoothness properties intrinsic to depth maps.Within the group sparse coding(GSC)framework,a combination of GBT and DGLR is implemented.Firstly,within each group,the graph is constructed by using estimates of the true values of the averaged blocks instead of the observations.Secondly,the graph Laplacian regular terms are constructed based on rows and columns of similar block groups,respectively.Lastly,the solution is obtained effectively by combining the alternating direction multiplication method(ADMM)with the weighted thresholding method within the domain of GBT.展开更多
Medical visual question answering(MedVQA)aims to enhance diagnostic confidence and deepen patientsunderstanding of their health conditions.While the Transformer architecture is widely used in multimodal fields,its app...Medical visual question answering(MedVQA)aims to enhance diagnostic confidence and deepen patientsunderstanding of their health conditions.While the Transformer architecture is widely used in multimodal fields,its application in MedVQA requires further enhancement.A critical limitation of contemporary MedVQA systems lies in the inability to integrate lifelong knowledge with specific patient data to generate human-like responses.Existing Transformer-based MedVQA models require enhancing their capabitities for interpreting answers through the applications of medical image knowledge.The introduction of the medical knowledge graph visual language transformer(MKGViLT),designed for joint medical knowledge graphs(KGs),addresses this challenge.MKGViLT incorporates an enhanced Transformer structure to effectively extract features and combine modalities for MedVQA tasks.The MKGViLT model delivers answers based on richer background knowledge,thereby enhancing performance.The efficacy of MKGViLT is evaluated using the SLAKE and P-VQA datasets.Experimental results show that MKGViLT surpasses the most advanced methods on the SLAKE dataset.展开更多
基金supported by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(RS-2025-00518960)in part by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(RS-2025-00563192).
文摘Mental-health risk detection seeks early signs of distress from social media posts and clinical transcripts to enable timely intervention before crises.When such risks go undetected,consequences can escalate to self-harm,long-term disability,reduced productivity,and significant societal and economic burden.Despite recent advances,detecting risk from online text remains challenging due to heterogeneous language,evolving semantics,and the sequential emergence of new datasets.Effective solutions must encode clinically meaningful cues,reason about causal relations,and adapt to new domains without forgetting prior knowledge.To address these challenges,this paper presents a Continual Neuro-Symbolic Graph Learning(CNSGL)framework that unifies symbolic reasoning,causal inference,and continual learning within a single architecture.Each post is represented as a symbolic graph linking clinically relevant tags to textual content,enriched with causal edges derived from directional Point-wise Mutual Information(PMI).A two-layer Graph Convolutional Network(GCN)encodes these graphs,and a Transformer-based attention pooler aggregates node embeddings while providing interpretable tag-level importances.Continual adaptation across datasets is achieved through the Multi-Head Freeze(MH-Freeze)strategy,which freezes a shared encoder and incrementally trains lightweight task-specific heads(small classifiers attached to the shared embedding).Experimental evaluations across six diverse mental-health datasets ranging from Reddit discourse to clinical interviews,demonstrate that MH-Freeze consistently outperforms existing continual-learning baselines in both discriminative accuracy and calibration reliability.Across six datasets,MH-Freeze achieves up to 0.925 accuracy and 0.923 F1-Score,with AUPRC≥0.934 and AUROC≥0.942,consistently surpassing all continual-learning baselines.The results confirm the framework’s ability to preserve prior knowledge,adapt to domain shifts,and maintain causal interpretability,establishing CNSGL as a promising step toward robust,explainable,and lifelong mental-health risk assessment.
基金supported by Basic Science Research Program through the National Research Foundation of Korea(NRF)funded by the Ministry of Education(RS-2023-00249743).
文摘This study proposes a novel forecasting framework that simultaneously captures the strong periodicity and irregular meteorological fluctuations inherent in solar radiation time series.Existing approaches typically define inter-regional correlations using either simple correlation coefficients or distance-based measures when applying spatio-temporal graph neural networks(STGNNs).However,such definitions are prone to generating spurious correlations due to the dominance of periodic structures.To address this limitation,we adopt the Elastic-Band Transform(EBT)to decompose solar radiation into periodic and amplitude-modulated components,which are then modeled independently with separate graph neural networks.The periodic component,characterized by strong nationwide correlations,is learned with a relatively simple architecture,whereas the amplitude-modulated component is modeled with more complex STGNNs that capture climatological similarities between regions.The predictions from the two components are subsequently recombined to yield final forecasts that integrate both periodic patterns and aperiodic variability.The proposed framework is validated with multiple STGNN architectures,and experimental results demonstrate improved predictive accuracy and interpretability compared to conventional methods.
基金supported in part by the National Natural Science Foundation of China(62372385).
文摘Dear Editor,This letter proposes the graph tensor alliance attention network(GT-A^(2)T)to represent a dynamic graph(DG)precisely.Its main idea includes 1)Establishing a unified spatio-temporal message propagation framework on a DG via the tensor product for capturing the complex cohesive spatio-temporal interdependencies precisely and 2)Acquiring the alliance attention scores by node features and favorable high-order structural correlations.
基金supported by the National Key Research and Development Program of China(No.2023YFF0905400)the National Natural Science Foundation of China(No.U2341229).
文摘Knowledge graphs(KGs),which organize real-world knowledge in triples,often suffer from issues of incompleteness.To address this,multi-hop knowledge graph reasoning(KGR)methods have been proposed for interpretable knowledge graph completion.The primary approaches to KGR can be broadly classified into two categories:reinforcement learning(RL)-based methods and sequence-to-sequence(seq2seq)-based methods.While each method has its own distinct advantages,they also come with inherent limitations.To leverage the strengths of each method while addressing their weaknesses,we propose a cyclical training method that alternates for several loops between the seq2seq training phase and the policy-based RL training phase using a transformer architecture.Additionally,a multimodal data encoding(MDE)module is introduced to improve the representation of entities and relations in KGs.TheMDE module treats entities and relations as distinct modalities,processing each with a dedicated network specialized for its respective modality.It then combines the representations of entities and relations in a dynamic and fine-grained manner using a gating mechanism.The experimental results from the knowledge graph completion task highlight the effectiveness of the proposed framework.Across five benchmark datasets,our framework achieves an average improvement of 1.7%in the Hits@1 metric and a 0.8%average increase in the Mean Reciprocal Rank(MRR)compared to other strong baseline methods.Notably,the maximum improvement in Hits@1 exceeds 4%,further demonstrating the effectiveness of the proposed approach.
基金the National Natural Science Foundation of China(Grant Nos.:62272288,U22A2041)Fundamental Research Funds for the Central Universities,Shaanxi Normal University(Grant No.:GK202302006)the Scientific Research Fund of Hunan Provincial Education Department of China(Grant No.:22B0097).
文摘Computational approaches for predicting drug-target interactions(DTIs)are pivotal in advancing drug discovery.Current methodologies leveraging heterogeneous networks often fall short in fully integrating both local and global network information.To comprehensively consider network information,we propose DHGT-DTI,a novel deep learning-based approach for DTI prediction.Specifically,we capture the local and global structural information of the network from both neighborhood and meta-path per-spectives.In the neighborhood perspective,we employ a heterogeneous graph neural network(HGNN),which extends Graph Sample and Aggregate(GraphSAGE)to handle diverse node and edge types,effectively learning local network structures.In the meta-path perspective,we introduce a Graph Transformer with residual connections to model higher-order relationships defined by meta-paths,such as"drug-disease-drug",and use an attention mechanism to fuse information across multiple meta-paths.The learned features from these dual perspectives are synergistically integrated for DTI prediction via a matrix decomposition method.Furthermore,DHGT-DTI reconstructs not only the DTI network but also auxiliary networks to bolster prediction accuracy.Comprehensive experiments on two benchmark datasets validate the superiority of DHGT-DTI over existing baseline methods.Additionally,case studies on six drugs used to treat Parkinson's disease not only validate the practical utility of DHGT-DTI but also highlight its broader potential in accelerating drug discovery for other diseases.
基金National Natural Science Foundation of China(No.62372100)。
文摘Owing to the constraints of depth sensing technology,images acquired by depth cameras are inevitably mixed with various noises.For depth maps presented in gray values,this research proposes a novel denoising model,termed graph-based transform(GBT)and dual graph Laplacian regularization(DGLR)(DGLR-GBT).This model specifically aims to remove Gaussian white noise by capitalizing on the nonlocal self-similarity(NSS)and the piecewise smoothness properties intrinsic to depth maps.Within the group sparse coding(GSC)framework,a combination of GBT and DGLR is implemented.Firstly,within each group,the graph is constructed by using estimates of the true values of the averaged blocks instead of the observations.Secondly,the graph Laplacian regular terms are constructed based on rows and columns of similar block groups,respectively.Lastly,the solution is obtained effectively by combining the alternating direction multiplication method(ADMM)with the weighted thresholding method within the domain of GBT.
基金Supported by the National Natural Science Foundation of China(No.62001313)the Liaoning Professional Talent Protect(No.XLYC2203046)the Shenyang Municipal Medical Engineering Cross Research Foundation of China(No.22-321-32-09).
文摘Medical visual question answering(MedVQA)aims to enhance diagnostic confidence and deepen patientsunderstanding of their health conditions.While the Transformer architecture is widely used in multimodal fields,its application in MedVQA requires further enhancement.A critical limitation of contemporary MedVQA systems lies in the inability to integrate lifelong knowledge with specific patient data to generate human-like responses.Existing Transformer-based MedVQA models require enhancing their capabitities for interpreting answers through the applications of medical image knowledge.The introduction of the medical knowledge graph visual language transformer(MKGViLT),designed for joint medical knowledge graphs(KGs),addresses this challenge.MKGViLT incorporates an enhanced Transformer structure to effectively extract features and combine modalities for MedVQA tasks.The MKGViLT model delivers answers based on richer background knowledge,thereby enhancing performance.The efficacy of MKGViLT is evaluated using the SLAKE and P-VQA datasets.Experimental results show that MKGViLT surpasses the most advanced methods on the SLAKE dataset.
