Pub Date : 2025-10-03DOI: 10.1109/TKDE.2025.3617894
Hao Wu;Qu Wang;Xin Luo;Zidong Wang
A nonstandard tensor is frequently adopted to model a large-sale complex dynamic network. A Tensor Representation Learning (TRL) model enables extracting valuable knowledge form a dynamic network via learning low-dimensional representation of a target nonstandard tensor. Nevertheless, the representation learning ability of existing TRL models are limited for a nonstandard tensor due to its inability to accurately represent the specific nature of the nonstandard tensor, i.e., mode imbalance, high-dimension, and incompleteness. To address this issue, this study innovatively proposes a Mode-Aware Tucker Network-based Tensor Representation Learning (MTN-TRL) model with three-fold ideas: a) designing a mode-aware Tucker network to accurately represent the imbalanced mode of a nonstandard tensor, b) building an MTN-based high-efficient TRL model that fuses both data density-oriented modeling principle and adaptive parameters learning scheme, and c) theoretically proving the MTN-TRL model’s convergence. Extensive experiments on eight nonstandard tensors generating from real-world dynamic networks demonstrate that MTN-TRL significantly outperforms state-of-the-art models in terms of representation accuracy.
{"title":"Learning Accurate Representation to Nonstandard Tensors via a Mode-Aware Tucker Network","authors":"Hao Wu;Qu Wang;Xin Luo;Zidong Wang","doi":"10.1109/TKDE.2025.3617894","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3617894","url":null,"abstract":"A nonstandard tensor is frequently adopted to model a large-sale complex dynamic network. A Tensor Representation Learning (TRL) model enables extracting valuable knowledge form a dynamic network via learning low-dimensional representation of a target nonstandard tensor. Nevertheless, the representation learning ability of existing TRL models are limited for a nonstandard tensor due to its inability to accurately represent the specific nature of the nonstandard tensor, i.e., mode imbalance, high-dimension, and incompleteness. To address this issue, this study innovatively proposes a Mode-Aware Tucker Network-based Tensor Representation Learning (MTN-TRL) model with three-fold ideas: a) designing a mode-aware Tucker network to accurately represent the imbalanced mode of a nonstandard tensor, b) building an MTN-based high-efficient TRL model that fuses both data density-oriented modeling principle and adaptive parameters learning scheme, and c) theoretically proving the MTN-TRL model’s convergence. Extensive experiments on eight nonstandard tensors generating from real-world dynamic networks demonstrate that MTN-TRL significantly outperforms state-of-the-art models in terms of representation accuracy.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7272-7285"},"PeriodicalIF":10.4,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455807","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-03DOI: 10.1109/TKDE.2025.3617461
Zhouyang Liu;Yixin Chen;Ning Liu;Jiezhong He;Dongsheng Li
Graph similarity is critical in graph-related tasks such as graph retrieval, where metrics like maximum common subgraph (MCS) and graph edit distance (GED) are commonly used. However, exact computations of these metrics are known to be NP-Hard. Recent neural network-based approaches approximate the similarity score in embedding spaces to alleviate the computational burden, but they either involve expensive pairwise node comparisons or fail to effectively utilize structural and scale information of graphs. To tackle these issues, we propose a novel geometric-based graph embedding method called Graph2Region (G2R). G2R represents nodes as closed regions and recovers their adjacency patterns within graphs in the embedding space. By incorporating the node features and adjacency patterns of graphs, G2R summarizes graph regions, i.e., graph embeddings, where the shape captures the underlying graph structures and the volume reflects the graph size. Consequently, the overlap between graph regions can serve as an approximation of MCS, signifying similar node regions and adjacency patterns. We further analyze the relationship between MCS and GED and propose using disjoint parts as a proxy for GED similarity. This analysis enables concurrent computation of MCS and GED, incorporating local and global structural information. Experimental evaluation highlights G2R’s competitive performance in graph similarity computation. It achieves up to a 60.0% relative accuracy improvement over state-of-the-art methods in MCS similarity learning, while maintaining efficiency in both training and inference. Moreover, G2R showcases remarkable capability in predicting both MCS and GED similarities simultaneously, providing a holistic assessment of graph similarity.
{"title":"Graph2Region: Efficient Graph Similarity Learning With Structure and Scale Restoration","authors":"Zhouyang Liu;Yixin Chen;Ning Liu;Jiezhong He;Dongsheng Li","doi":"10.1109/TKDE.2025.3617461","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3617461","url":null,"abstract":"Graph similarity is critical in graph-related tasks such as graph retrieval, where metrics like maximum common subgraph (MCS) and graph edit distance (GED) are commonly used. However, exact computations of these metrics are known to be NP-Hard. Recent neural network-based approaches approximate the similarity score in embedding spaces to alleviate the computational burden, but they either involve expensive pairwise node comparisons or fail to effectively utilize structural and scale information of graphs. To tackle these issues, we propose a novel geometric-based graph embedding method called <sc>Graph2Region</small> (<sc>G2R</small>). <sc>G2R</small> represents nodes as closed regions and recovers their adjacency patterns within graphs in the embedding space. By incorporating the node features and adjacency patterns of graphs, <sc>G2R</small> summarizes graph regions, i.e., graph embeddings, where the shape captures the underlying graph structures and the volume reflects the graph size. Consequently, the overlap between graph regions can serve as an approximation of MCS, signifying similar node regions and adjacency patterns. We further analyze the relationship between MCS and GED and propose using disjoint parts as a proxy for GED similarity. This analysis enables concurrent computation of MCS and GED, incorporating local and global structural information. Experimental evaluation highlights <sc>G2R</small>’s competitive performance in graph similarity computation. It achieves up to a 60.0% relative accuracy improvement over state-of-the-art methods in MCS similarity learning, while maintaining efficiency in both training and inference. Moreover, <sc>G2R</small> showcases remarkable capability in predicting both MCS and GED similarities simultaneously, providing a holistic assessment of graph similarity.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7213-7225"},"PeriodicalIF":10.4,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455944","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Data privacy protection legislation around the world has increasingly enforced the “right to be forgotten” regulation, generating a surge in research interest in machine unlearning (MU), which aims to remove the impact of training data from machine learning models upon receiving revocation requests from data owners. There exist two major challenges for the performance of MU: the execution efficiency and the inference interference. The former requires minimizing the computational overhead for each execution of the MU mechanism, while the latter calls for reducing the execution frequency to minimize interference with normal inference services. Nowadays most MU studies focus on the sample-level unlearning setting, leaving the other paramount feature-level setting under-explored. Adapting these existing techniques to the latter turns out to be non-trivial. The only known feature-level work achieves an approximate unlearning guarantee, but suffers from degraded model accuracy and still leaves the inference interference challenge unsolved. We are therefore motivated to propose FELEMN, the first FEature-Level Exact Machine uNlearning method that overcomes both of the above-mentioned hurdles. For the MU execution efficiency challenge, we explore the impact of different feature partitioning strategies on the preservation of semantic relationships for maintaining model accuracy and MU efficiency. For the inference interference challenge, we propose two batching mechanisms to combine as many individual unlearning requests to be processed together as possible, while avoiding potential privacy issues coming with falsely postponing unlearning requests, which is grounded on theoretical analysis. Experiments on five real datasets show that our FELEMN outperforms up-to-date competitors with up to $3times$ speedup for each MU execution, and 50% runtime reduction by mitigating inference interference.
{"title":"FELEMN: Toward Efficient Feature-Level Machine Unlearning for Exact Privacy Protection","authors":"Zhigang Wang;Yizhen Yu;Mingxin Li;Jian Lou;Ning Wang;Yu Gu;Shen Su;Yuan Liu;Hui Jiang;Zhihong Tian","doi":"10.1109/TKDE.2025.3613659","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3613659","url":null,"abstract":"Data privacy protection legislation around the world has increasingly enforced the “right to be forgotten” regulation, generating a surge in research interest in machine unlearning (MU), which aims to remove the impact of training data from machine learning models upon receiving revocation requests from data owners. There exist two major challenges for the performance of MU: the execution efficiency and the inference interference. The former requires minimizing the computational overhead for each execution of the MU mechanism, while the latter calls for reducing the execution frequency to minimize interference with normal inference services. Nowadays most MU studies focus on the sample-level unlearning setting, leaving the other paramount feature-level setting under-explored. Adapting these existing techniques to the latter turns out to be non-trivial. The only known feature-level work achieves an <i>approximate</i> unlearning guarantee, but suffers from degraded model accuracy and still leaves the inference interference challenge unsolved. We are therefore motivated to propose FELEMN, the first FEature-Level Exact Machine uNlearning method that overcomes both of the above-mentioned hurdles. For the MU execution efficiency challenge, we explore the impact of different feature partitioning strategies on the preservation of semantic relationships for maintaining model accuracy and MU efficiency. For the inference interference challenge, we propose two batching mechanisms to combine as many individual unlearning requests to be processed together as possible, while avoiding potential privacy issues coming with falsely postponing unlearning requests, which is grounded on theoretical analysis. Experiments on five real datasets show that our FELEMN outperforms up-to-date competitors with up to <inline-formula><tex-math>$3times$</tex-math></inline-formula> speedup for each MU execution, and 50% runtime reduction by mitigating inference interference.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7169-7183"},"PeriodicalIF":10.4,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145456019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Existing cold-start recommendation methods typically use item-level alignment strategies to align the content feature and collaborative feature of warm items during model training. However, these methods are less effective for cold items with low semantic similarity to the warm items when they first appear in the test stage, as they have no historical interactions to obtain the collaborative feature. In this paper, we propose a preference aware recommendation (PARec) model with hierarchical item alignment to solve the item cold-start issue. Our approach exploits user preference from historical records to achieve group-level alignment with item content feature, enhancing recommendation performance. Specifically, our hierarchical item alignment strategy improves recommendations for both high and low similarity cold items by using item-level alignment for high similarity cold items and introducing group-level alignment for low similarity cold items. Low similarity cold items can be successfully recommended through relationships among items, captured by our group-level alignment, based on their co-occurrence possibilities and semantic similarities. For model training, a hierarchical contrastive objective function is presented to balance the performance of warm and cold items, achieving better overall performance. Extensive experiments demonstrate the effectiveness of our method, with results showing its superiority compared to state-of-the-art approaches.
{"title":"Preference Aware Item Cold-Start Recommendation With Hierarchical Item Alignment","authors":"Wenbo Wang;Ben Chen;Bingquan Liu;Lili Shan;Chengjie Sun;Qian Chen;Feiyang Xiao;Jian Guan","doi":"10.1109/TKDE.2025.3613263","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3613263","url":null,"abstract":"Existing cold-start recommendation methods typically use item-level alignment strategies to align the content feature and collaborative feature of warm items during model training. However, these methods are less effective for cold items with low semantic similarity to the warm items when they first appear in the test stage, as they have no historical interactions to obtain the collaborative feature. In this paper, we propose a preference aware recommendation (PARec) model with hierarchical item alignment to solve the item cold-start issue. Our approach exploits user preference from historical records to achieve group-level alignment with item content feature, enhancing recommendation performance. Specifically, our hierarchical item alignment strategy improves recommendations for both high and low similarity cold items by using item-level alignment for high similarity cold items and introducing group-level alignment for low similarity cold items. Low similarity cold items can be successfully recommended through relationships among items, captured by our group-level alignment, based on their co-occurrence possibilities and semantic similarities. For model training, a hierarchical contrastive objective function is presented to balance the performance of warm and cold items, achieving better overall performance. Extensive experiments demonstrate the effectiveness of our method, with results showing its superiority compared to state-of-the-art approaches.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7388-7401"},"PeriodicalIF":10.4,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145456052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Visual Question Answering (VQA), aimed at improving AI-driven interactions and solving complex visual-linguistic tasks, has increasingly garnered attention as a pivotal research domain in both academic and industrial spheres. Despite progress in VQA, current studies still suffer from the challenge of language bias posed by spurious semantic correlations and minority class collapse, leading to semantic ambiguities and distribution shifts that hinder robust performance across challenging scenarios. To address these challenges, we propose a robust multi-space collaborative debiasing paradigm, termed “LBF-VQA”, which systematically leverages multi-space collaborative debiasing strategies to achieve language bias-free VQA, encompassing both Euclidean space debiasing (ESD) and Spherical space debiasing (SSD). By strategically introducing bias-examples and their corresponding counter-examples, the ESD strategy focuses on uncovering hidden prior correlations and the complex interactions between modality and semantics within the Euclidean space. Benefiting from the infinite contrastive and distribution debiasing learning mechanisms, the SSD strategy is devoted to effectively preventing the collapse of minority classes while enhancing the manifold representations of instance de-bias and distribution de-dependence in the Spherical space. Furthermore, we meticulously constructed a specialized medical dataset intentionally embedded with deliberate language bias to comprehensively examine the negative effects of language bias on medical VQA systems. Extensive experiments on multiple general and medical VQA benchmarks consistently verify the effectiveness and generalizability of our LBF-VQA in handling various complex VQA scenarios than state-of-the-art baselines.
{"title":"LBF-VQA: Towards Language Bias-Free Visual Question Answering With Multi-Space Collaborative Debiasing","authors":"Yishu Liu;Huanjia Zhu;Bingzhi Chen;Xiaozhao Fang;Guangming Lu;Shengli Xie","doi":"10.1109/TKDE.2025.3613421","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3613421","url":null,"abstract":"Visual Question Answering (VQA), aimed at improving AI-driven interactions and solving complex visual-linguistic tasks, has increasingly garnered attention as a pivotal research domain in both academic and industrial spheres. Despite progress in VQA, current studies still suffer from the challenge of <italic>language bias</i> posed by spurious semantic correlations and minority class collapse, leading to semantic ambiguities and distribution shifts that hinder robust performance across challenging scenarios. To address these challenges, we propose a robust multi-space collaborative debiasing paradigm, termed “LBF-VQA”, which systematically leverages multi-space collaborative debiasing strategies to achieve language bias-free VQA, encompassing both Euclidean space debiasing (ESD) and Spherical space debiasing (SSD). By strategically introducing bias-examples and their corresponding counter-examples, the ESD strategy focuses on uncovering hidden prior correlations and the complex interactions between modality and semantics within the Euclidean space. Benefiting from the infinite contrastive and distribution debiasing learning mechanisms, the SSD strategy is devoted to effectively preventing the collapse of minority classes while enhancing the manifold representations of instance de-bias and distribution de-dependence in the Spherical space. Furthermore, we meticulously constructed a specialized medical dataset intentionally embedded with deliberate language bias to comprehensively examine the negative effects of language bias on medical VQA systems. Extensive experiments on multiple general and medical VQA benchmarks consistently verify the effectiveness and generalizability of our LBF-VQA in handling various complex VQA scenarios than state-of-the-art baselines.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7255-7271"},"PeriodicalIF":10.4,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455968","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-22DOI: 10.1109/TKDE.2025.3607765
Yaming Yang;Zhe Wang;Ziyu Guan;Wei Zhao;Xinyan Huang;Xiaofei He
Entity Alignment (EA) is to link potential equivalent entities across different knowledge graphs (KGs). Most existing EA methods are supervised as they require the supervision of seed alignments, i.e., manually specified aligned entity pairs. Very recently, several EA studies have made some attempts to get rid of seed alignments. Despite achieving preliminary progress, they still suffer two limitations: (1) The entity embeddings produced by their GNN-like encoders lack personalization since some of the aggregation subpaths are shared between different entities. (2) They cannot fully alleviate the distribution distortion issue between candidate KGs due to the absence of supervised signals. In this work, we propose a novel unsupervised entity alignment approach called UNEA to address the above two issues. First, we parametrically sample a tree neighborhood rooted at each entity, and accordingly develop a tree attention aggregation mechanism to extract a personalized embedding for each entity. Second, we introduce an auxiliary task of maximizing the mutual information between the input and the output of the KG encoder, which serves as a regularization to prevent the distribution distortion. Extensive experiments show that our UNEA achieves a new state-of-the-art for the unsupervised EA task, and can even outperform many existing supervised EA baselines.
{"title":"Unsupervised Entity Alignment Based on Personalized Discriminative Rooted Tree","authors":"Yaming Yang;Zhe Wang;Ziyu Guan;Wei Zhao;Xinyan Huang;Xiaofei He","doi":"10.1109/TKDE.2025.3607765","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3607765","url":null,"abstract":"Entity Alignment (EA) is to link potential equivalent entities across different knowledge graphs (KGs). Most existing EA methods are supervised as they require the supervision of seed alignments, i.e., manually specified aligned entity pairs. Very recently, several EA studies have made some attempts to get rid of seed alignments. Despite achieving preliminary progress, they still suffer two limitations: (1) The entity embeddings produced by their GNN-like encoders lack personalization since some of the aggregation subpaths are shared between different entities. (2) They cannot fully alleviate the distribution distortion issue between candidate KGs due to the absence of supervised signals. In this work, we propose a novel unsupervised entity alignment approach called UNEA to address the above two issues. First, we parametrically sample a tree neighborhood rooted at each entity, and accordingly develop a tree attention aggregation mechanism to extract a personalized embedding for each entity. Second, we introduce an auxiliary task of maximizing the mutual information between the input and the output of the KG encoder, which serves as a regularization to prevent the distribution distortion. Extensive experiments show that our UNEA achieves a new state-of-the-art for the unsupervised EA task, and can even outperform many existing supervised EA baselines.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7440-7452"},"PeriodicalIF":10.4,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455963","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-22DOI: 10.1109/TKDE.2025.3613148
Hao Huang;Mingxin Wang;Mengqi Shan;Zhigao Zheng;Ting Gan;Jiawei Jiang;Zongpeng Li
Outdoor billboard advertising has proven effective for commercial promotions, attracting potential customers, and boosting product sales. Auction serves as a popular method for leasing billboard usage rights, enabling a seller to rent billboards to winning users for predefined periods according to their bids. An effective auction algorithm is of great significance to maximize the efficiency of the billboard ecosystem. In contrast to a rich literature on Internet advertising auctions, well-crafted algorithms tailored for outdoor billboard auctions remain rare. In this work, we investigate the problem of outdoor billboard auctions, in the practical setting where bids are received and processed on the fly. Our goal is to maximize social welfare, namely the total benefits of auction participants, including the billboard service provider and the bidding users. To this end, we first formulate the billboard social welfare maximization problem into an Integer Linear Problem (ILP), and then reformulate the ILP into a compact form with a reduced size of constraints (at the cost of involving exponentially many primal variables), based on which we derive the dual problem. Furthermore, we design a dual oracle to handle the exponentially many dual constraints, avoiding exhaustive enumeration. We present a primal-dual online algorithm with an incentive-compatible pricing mechanism. Theoretical analysis proves the individual rationality, incentive compatibility, and computational efficiency of our online algorithm. Extensive experimental results show that the online algorithm is both effective and efficient, and achieves a good competitive ratio.
{"title":"Online Billboard Auction With Social Welfare Maximization","authors":"Hao Huang;Mingxin Wang;Mengqi Shan;Zhigao Zheng;Ting Gan;Jiawei Jiang;Zongpeng Li","doi":"10.1109/TKDE.2025.3613148","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3613148","url":null,"abstract":"Outdoor billboard advertising has proven effective for commercial promotions, attracting potential customers, and boosting product sales. Auction serves as a popular method for leasing billboard usage rights, enabling a seller to rent billboards to winning users for predefined periods according to their bids. An effective auction algorithm is of great significance to maximize the efficiency of the billboard ecosystem. In contrast to a rich literature on Internet advertising auctions, well-crafted algorithms tailored for outdoor billboard auctions remain rare. In this work, we investigate the problem of outdoor billboard auctions, in the practical setting where bids are received and processed on the fly. Our goal is to maximize social welfare, namely the total benefits of auction participants, including the billboard service provider and the bidding users. To this end, we first formulate the billboard social welfare maximization problem into an Integer Linear Problem (ILP), and then reformulate the ILP into a compact form with a reduced size of constraints (at the cost of involving exponentially many primal variables), based on which we derive the dual problem. Furthermore, we design a dual oracle to handle the exponentially many dual constraints, avoiding exhaustive enumeration. We present a primal-dual online algorithm with an incentive-compatible pricing mechanism. Theoretical analysis proves the individual rationality, incentive compatibility, and computational efficiency of our online algorithm. Extensive experimental results show that the online algorithm is both effective and efficient, and achieves a good competitive ratio.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7362-7373"},"PeriodicalIF":10.4,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145456020","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The demand for more precise and timely urban resource allocation and management has driven the extension of urban flow prediction from short-term to long-term horizons. As the time scale expands, the issue of urban flow distribution shift becomes increasingly prominent due to various impact factors, such as weather, events, city changes, etc. Traditionally, comprehensively analyzing and addressing the causal relationships underlying the distribution shift caused by these factors has been challenging. In this paper, we propose that these impact factors can be partitioned in two major types, i.e., context factors and structural factors. We then present a decomposition-based model for long-term urban flow prediction from a causal perspective, named DeCau, which can discriminate between the two types of factors for effectively solving the problem of urban flow distribution shift. First, we employ a decomposition module to decompose urban flow into seasonal part and trend part. The seasonal part contains high frequency irregular variations caused by context factors. We advise a shared distribution estimator to approximate the unavailable prior distributions of context factors, and then apply causal intervention to mitigate the confounding impact of context factors. The distribution shift in the trend part is induced by structural factors. We design a dual causal dependency extractor to model the causality between POIs distribution and urban flow, and then eliminate spurious correlations through causal adjustment. Finally, we design an end-to-end framework for long-term urban flow prediction by combining the embeddings from two parts, enabling the model to generalize to unseen distribution. Extensive experimental results demonstrate DeCau outperforms state-of-the-art baselines.
{"title":"Long-Term Urban Flow Prediction Against Data Distribution Shift: A Causal Perspective","authors":"Yuting Liu;Qiang Zhou;Hanzhe Li;Fuzhen Zhuang;Jingjing Gu","doi":"10.1109/TKDE.2025.3612033","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3612033","url":null,"abstract":"The demand for more precise and timely urban resource allocation and management has driven the extension of urban flow prediction from short-term to long-term horizons. As the time scale expands, the issue of urban flow distribution shift becomes increasingly prominent due to various impact factors, such as weather, events, city changes, etc. Traditionally, comprehensively analyzing and addressing the causal relationships underlying the distribution shift caused by these factors has been challenging. In this paper, we propose that these impact factors can be partitioned in two major types, i.e., context factors and structural factors. We then present a decomposition-based model for long-term urban flow prediction from a causal perspective, named <italic>DeCau</i>, which can discriminate between the two types of factors for effectively solving the problem of urban flow distribution shift. First, we employ a decomposition module to decompose urban flow into seasonal part and trend part. The seasonal part contains high frequency irregular variations caused by context factors. We advise a shared distribution estimator to approximate the unavailable prior distributions of context factors, and then apply causal intervention to mitigate the confounding impact of context factors. The distribution shift in the trend part is induced by structural factors. We design a dual causal dependency extractor to model the causality between POIs distribution and urban flow, and then eliminate spurious correlations through causal adjustment. Finally, we design an end-to-end framework for long-term urban flow prediction by combining the embeddings from two parts, enabling the model to generalize to unseen distribution. Extensive experimental results demonstrate <italic>DeCau</i> outperforms state-of-the-art baselines.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7286-7299"},"PeriodicalIF":10.4,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455973","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Bipartite graphs are widely used in many real-world applications, where discovering clusters is crucial for understanding their underlying structure. However, most existing clustering methods for bipartite graphs enforce the assignment of all vertices to clusters, often neglecting the important roles of outliers and hubs. To address this limitation, we plan to extend the structural clustering model from unipartite to bipartite graphs. This extension is non-trivial due to the lack of common neighbors in bipartite graphs, which renders traditional similarity measures less effective. Recognizing that similarity is key to structural clustering, we resort to butterflies—the fundamental building blocks of bipartite graphs—to define a more effective similarity measure. Building on this, we further propose a novel structural clustering model, ${mathsf {SBC}}$, tailored for bipartite graphs. To enable clustering under this model, we develop efficient online and index-based methods, along with a dynamic maintenance method to accommodate graph updates over time. Extensive experiments on real-world bipartite graphs demonstrate that: (1) The ${mathsf {SBC}}$ model greatly enhances clustering quality, achieving higher modularity while effectively identifying outliers and hubs. (2) Our proposed clustering methods are highly scalable, enabling the processing of graphs with up to 12.2 million edges within 2 seconds.
{"title":"Structural Clustering for Bipartite Graphs","authors":"Mingyu Yang;Wentao Li;Wei Wang;Dong Wen;Min Gao;Lu Qin","doi":"10.1109/TKDE.2025.3612290","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3612290","url":null,"abstract":"Bipartite graphs are widely used in many real-world applications, where discovering clusters is crucial for understanding their underlying structure. However, most existing clustering methods for bipartite graphs enforce the assignment of <i>all</i> vertices to clusters, often neglecting the important roles of outliers and hubs. To address this limitation, we plan to extend the structural clustering model from unipartite to bipartite graphs. This extension is non-trivial due to the lack of common neighbors in bipartite graphs, which renders traditional similarity measures less effective. Recognizing that similarity is key to structural clustering, we resort to butterflies—the fundamental building blocks of bipartite graphs—to define a more effective similarity measure. Building on this, we further propose a novel structural clustering model, <inline-formula><tex-math>${mathsf {SBC}}$</tex-math></inline-formula>, tailored for bipartite graphs. To enable clustering under this model, we develop efficient online and index-based methods, along with a dynamic maintenance method to accommodate graph updates over time. Extensive experiments on real-world bipartite graphs demonstrate that: (1) The <inline-formula><tex-math>${mathsf {SBC}}$</tex-math></inline-formula> model greatly enhances clustering quality, achieving higher modularity while effectively identifying outliers and hubs. (2) Our proposed clustering methods are highly scalable, enabling the processing of graphs with up to 12.2 million edges within 2 seconds.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"38 1","pages":"645-658"},"PeriodicalIF":10.4,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705912","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Recently, large language models (LLMs) have made remarkable progress in table understanding, yet they remain vulnerable to the structural noise (SN) and the textual noise (TN). Existing methods usually employ biased denoising strategies such as structural matching and textual filtering, or overzealous denoising strategies such as introducing supplementary tasks like text-to-SQL and table-to-text to reduce these two types of noise. However, these methods either neglect one type of noise or introduce substantial external noise. Therefore, how to simultaneously mitigate the structural and textual noise without introducing extra noise and improve the performance of LLMs in table understanding is still an unresolved issue. In this paper, we rethink the bottlenecks in table understanding from the perspective of noise reduction and propose a novel dual-denoiser-reasoner model, called TabDDR, for balanced and effective denoising. Specially, our model consists of a structural-and-textual denoiser and a task-adaptive reasoner. The former removes two types of noise via triplet alignment and planning extraction to seek an interpretable balance between breaking structural barriers and preserving structural characteristics, eliminating textual noise and retaining maximal information; the latter ensures a simple but effective reasoning process which can adapt to various downstream tasks. To highlight the presence and impact of the structural and textual noise, we construct the WTQ-SN and WTQ-TN datasets based on the WikiTableQuestion (WTQ) dataset. Extensive experiments on these self-constructed datasets and two other public datasets demonstrate that our proposed method performs better than state-of-the-art baselines.
{"title":"Toward Balanced Denoising: Building a Structural and Textual Denoiser for Table Understanding","authors":"Shu-Xun Yang;Xian-Ling Mao;Yu-Ming Shang;Heyan Huang","doi":"10.1109/TKDE.2025.3612217","DOIUrl":"https://doi.org/10.1109/TKDE.2025.3612217","url":null,"abstract":"Recently, large language models (LLMs) have made remarkable progress in table understanding, yet they remain vulnerable to the structural noise (SN) and the textual noise (TN). Existing methods usually employ biased denoising strategies such as structural matching and textual filtering, or overzealous denoising strategies such as introducing supplementary tasks like text-to-SQL and table-to-text to reduce these two types of noise. However, these methods either neglect one type of noise or introduce substantial external noise. Therefore, how to simultaneously mitigate the structural and textual noise without introducing extra noise and improve the performance of LLMs in table understanding is still an unresolved issue. In this paper, we rethink the bottlenecks in table understanding from the perspective of noise reduction and propose a novel dual-denoiser-reasoner model, called TabDDR, for balanced and effective denoising. Specially, our model consists of a structural-and-textual denoiser and a task-adaptive reasoner. The former removes two types of noise via triplet alignment and planning extraction to seek an interpretable balance between breaking structural barriers and preserving structural characteristics, eliminating textual noise and retaining maximal information; the latter ensures a simple but effective reasoning process which can adapt to various downstream tasks. To highlight the presence and impact of the structural and textual noise, we construct the WTQ-SN and WTQ-TN datasets based on the WikiTableQuestion (WTQ) dataset. Extensive experiments on these self-constructed datasets and two other public datasets demonstrate that our proposed method performs better than state-of-the-art baselines.","PeriodicalId":13496,"journal":{"name":"IEEE Transactions on Knowledge and Data Engineering","volume":"37 12","pages":"7414-7425"},"PeriodicalIF":10.4,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145456048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：