{"title":"Distributed Multi-Task Learning for Stochastic Bandits With Context Distribution and Stage-Wise Constraints","authors":"Jiabin Lin;Shana Moothedath","doi":"10.1109/TSIPN.2025.3566239","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3566239","url":null,"abstract":"We present conservative multi-task learning in stochastic linear contextual bandits with <italic>heterogeneous</i> agents. This extends conservative linear bandits to a distributed setting where <inline-formula><tex-math>$M$</tex-math></inline-formula> agents tackle <italic>different but related</i> tasks while adhering to stage-wise performance constraints. The exact context is <italic>unknown</i>, and only a context distribution is available to the agents as in many practical applications that involve a prediction mechanism to infer context, such as stock market prediction and weather forecast. We propose a distributed upper confidence bound (UCB) algorithm, DiSC-UCB. Our algorithm dynamically constructs a pruned action set for each task in every round, guaranteeing compliance with the constraints. Additionally, it includes synchronized sharing of estimates among agents via a central server using well-structured synchronization steps. For <inline-formula><tex-math>$d$</tex-math></inline-formula>-dimensional linear bandits, we prove an <inline-formula><tex-math>$widetilde{O}(dsqrt{MT})$</tex-math></inline-formula> regret bound and an <inline-formula><tex-math>$O(M^{1.5}d^{3})$</tex-math></inline-formula> communication bound on the algorithm. We extend the problem to a setting where the agents are unaware of the baseline reward. We provide a modified algorithm, DiSC-UCB-UB, and show that it achieves the same regret and communication bounds. We empirically validated the performance of our algorithm on synthetic data and real-world Movielens-100 K and LastFM data and also compared it with some existing benchmark algorithms.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"577-591"},"PeriodicalIF":3.0,"publicationDate":"2025-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524397","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Distributed Adaptive Signal Fusion for Fractional Programs","authors":"Cem Ates Musluoglu;Alexander Bertrand","doi":"10.1109/TSIPN.2025.3546462","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3546462","url":null,"abstract":"The distributed adaptive signal fusion (DASF) is an algorithmic framework that allows solving spatial filtering optimization problems in a distributed and adaptive fashion over a bandwidth-constrained wireless sensor network. The DASF algorithm requires each node to sequentially build a compressed version of the original network-wide problem and solve it locally. However, these local problems can still result in a high computational load at the nodes, especially when the required solver is iterative. In this paper, we study the particular case of fractional programs, i.e., problems for which the objective function is a fraction of two continuous functions, which indeed require such iterative solvers. By exploiting the structure of a commonly used method for solving fractional programs and interleaving it with the iterations of the standard DASF algorithm, we obtain a distributed algorithm with a significantly reduced computational cost compared to the straightforward application of DASF as a meta-algorithm. We prove convergence and optimality of this “fractional DASF” (F-DASF) algorithm and demonstrate its performance via numerical simulations.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"258-273"},"PeriodicalIF":3.0,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143637844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Peng Cai;Dongyuan Lin;Junhui Qian;Yunfei Zheng;Zhongyuan Guo;Shiyuan Wang
{"title":"Distributed Cascaded Spline-Based Adaptive Graph Filters for Nonlinear Systems: Design and Performance Analysis","authors":"Peng Cai;Dongyuan Lin;Junhui Qian;Yunfei Zheng;Zhongyuan Guo;Shiyuan Wang","doi":"10.1109/TSIPN.2025.3546469","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3546469","url":null,"abstract":"The distributed nonlinear adaptive graph filter (DNAGF) is developed with the single nonlinear graph filter model (NGFM) to handle streaming datasets. However, the current DNAGFs tend to underperform when predicting unknown nonlinear dynamic systems. This suboptimal performance is due to their reliance on a single NGFM and the network's limited computational burden. To address these issues, two novel cascaded DNAGFs considering the spline interpolation method, i.e. a distributed Wiener spline adaptive graph filter (DWSAGF) and distributed Hammerstein spline adaptive graph filter (DHSAGF), are proposed to improve the capacity for nonlineaprediction in this paper. By utilizing piecewise low-order nonlinear spline functions, the proposed DWSAGF and DHSAGF can adapt locally to improve the fitting of the predicted nonlinear system to the unknown one. In DWSAGF and DHSAGF, the cascaded architectures containing linear and nonlinear subsystems are employed, which are more flexible than the single NGFM. Particularly, since DHSAGF has a memory through the constructed matrix <inline-formula><tex-math>${boldsymbol {bar{U}}}_{m}^{bar{t}}(r)$</tex-math></inline-formula>, it generates higher performance than DWSAGF for complex or time-varying nonlinear systems. In addition, the detailed performance analyses regarding DWSAGF and DHSAGF in the mean and mean-square senses are presented. Simulations are exhibited to validate the theoretical analysis and to show the performance superiorities of DWSAGF and DHSAGF.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"274-288"},"PeriodicalIF":3.0,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143637851","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Distributed Online Learning Over Multitask Networks With Rank-One Model","authors":"Yitong Chen;Danqi Jin;Jie Chen;Cédric Richard;Wen Zhang;Gongping Huang;Jingdong Chen","doi":"10.1109/TSIPN.2025.3543973","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3543973","url":null,"abstract":"Modeling multitask relations in distributed networks has garnered considerable interest in recent years. In this paper, we present a novel rank-one model, where all the optimal vectors to be estimated are scaled versions of an unknown vector to be determined. By considering the rank-one relation, we develop a constrained centralized optimization problem, and after a decoupling process, it is solved in a distributed way by using the projected gradient descent method. To perform an efficient calculation of this projection, we suggest substituting the intensive singular value decomposition with the computationally efficient power method. Additionally, local estimates targeting the same optimal vector are combined within a neighborhood to further improve their accuracy. Theoretical analyses of the proposed algorithm are conducted for star topologies, and conditions are derived to guarantee its stability in both the mean and mean-square senses. Finally, simulation results are presented to demonstrate the effectiveness of the proposed algorithms.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"314-328"},"PeriodicalIF":3.0,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143675974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Adaptive Joint Estimation of Temporal Vertex and Edge Signals","authors":"Yi Yan;Tian Xie;Ercan E. Kuruoglu","doi":"10.1109/TSIPN.2025.3536084","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3536084","url":null,"abstract":"The adaptive estimation of coexisting temporal vertex (node) and edge signals on graphs is a critical task when a change in edge signals influences the temporal dynamics of the vertex signals. Existing Graph Signal Processing algorithms have extensively studied signals on graph vertices, and while recent advancements have started exploring signals on edges, a framework for systematically representing interactive time-varying signals across vertices, edges, and higher-order structures has yet to be fully realized. We propose an Adaptive Joint Vertex-Edge Estimation (AJVEE) algorithm for jointly estimating time-varying vertex and edge signals through a time-varying regression, incorporating both vertex signal filtering and edge signal filtering. Accompanying AJVEE is a newly proposed Adaptive Least Mean Square procedure based on the Hodge Laplacian (ALMS-Hodge), which is inspired by classical adaptive filters combining simplicial filtering and simplicial regression. AJVEE is able to operate jointly on the vertices and edges by merging two ALMS-Hodge specified on the vertices and edges into a unified formulation. A more generalized case extending AJVEE beyond the vertices and edges is being discussed. Experimenting on real-world traffic networks and population mobility networks, we have confirmed that our proposed AJVEE algorithm could accurately and jointly track time-varying vertex and edge signals on graphs.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"215-229"},"PeriodicalIF":3.0,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143553483","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zirui Liao;Jian Shi;Shaoping Wang;Yuwei Zhang;Rentong Chen;Zhiyong Sun
{"title":"Dynamic Event-Triggering Resilient Coordination for Time-Varying Heterogeneous Networks","authors":"Zirui Liao;Jian Shi;Shaoping Wang;Yuwei Zhang;Rentong Chen;Zhiyong Sun","doi":"10.1109/TSIPN.2025.3541932","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3541932","url":null,"abstract":"This study addresses the resilient coordination problem for heterogeneous multi-agent systems (MASs) consisting of first-order and second-order agents in time-invariant and time-varying networks. An internal dynamic variable is introduced to flexibly adjust the triggering threshold and facilitate the dynamic event-triggering condition (DETC). Under adversarial attacks, a novel resilient consensus strategy called <italic>heterogeneous dynamic event-triggering mean-subsequence-reduced (HDE-MSR) algorithm</i> is further developed, which ensures that the positions of all healthy agents achieve consensus on the identical value and the velocities of all healthy second-order agents asymptotically approach zero despite the influence of faulty agents. Moreover, the resilient consensus in time-varying networks is further guaranteed by the introduction of jointly robust graphs. Finally, three case studies are provided to validate the effectiveness and superior performance of the HDE-MSR algorithm.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"329-341"},"PeriodicalIF":3.0,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143667269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"The Graph Fractional Fourier Transform in Hilbert Space","authors":"Yu Zhang;Bing-Zhao Li","doi":"10.1109/TSIPN.2025.3540714","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3540714","url":null,"abstract":"Graph signal processing (GSP) leverages the inherent signal structure within graphs to extract high-dimensional data without relying on translation invariance. It has emerged as a crucial tool across multiple fields, including learning and processing of various networks, data analysis, and image processing. In this paper, we introduce the graph fractional Fourier transform in Hilbert space (HGFRFT), which provides additional fractional analysis tools for generalized GSP by extending Hilbert space and vertex domain Fourier analysis to fractional order. First, we establish that the proposed HGFRFT extends traditional GSP, accommodates graphs on continuous domains, and facilitates joint time-vertex domain transform while adhering to critical properties such as additivity, commutativity, and invertibility. Second, to process generalized graph signals in the fractional domain, we explore the theory behind filtering and sampling of signals in the fractional domain. Finally, our simulations and numerical experiments substantiate the advantages and enhancements yielded by the HGFRFT.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"242-257"},"PeriodicalIF":3.0,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143601899","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Learning General Brain Network Representations of Different Brain Disorders Using Invariant Subgraph GNN","authors":"Hao Zhang;Ran Song;Liping Wang;Lei Mou;Yushan Lu;Yitian Zhao;Wei Zhang","doi":"10.1109/TSIPN.2025.3540709","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3540709","url":null,"abstract":"Distribution shifts across data from various brain disorders pose significant challenges for diagnosis. Establishing general feature representations that can handle these distribution shifts is crucial for accurately diagnosing these conditions. However, this area remains largely unexplored. This work propose an Invariant Subgraph GNN (IS-GNN) to learn general brain network representations for classifying various brain disorders in resting-state fMRI. This model employs an invariant subgraph learning mechanism to capture invariant brain graphs and handle distribution shifts. Moreover, we have developed an adaptive structure perception module to improve the detection of invariant subgraph features in brain networks by assessing the importance of nodes within the brain graph. To further refine the model, we propose a self-supervised loss for invariant subgraph learning, ensuring the generation of invariant brain network representations. Pretrained on data from 1,943 subjects across three public datasets corresponding to Autism Spectrum Disorder, Attention Deficit Hyperactivity Disorder, and Parkinson's Disease, the fine-tuning experiments of our proposed method demonstrate that the model achieves the state-of-the-art classification performance on not only the three datasets but also on an external Alzheimer's Disease dataset across.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"230-241"},"PeriodicalIF":3.0,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143594439","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Community Detection From Multiple Observations: From Product Graph Model to Brain Applications","authors":"Tiziana Cattai;Gaetano Scarano;Marie-Constance Corsi;Fabrizio DeVico Fallani;Stefania Colonnese","doi":"10.1109/TSIPN.2025.3540702","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3540702","url":null,"abstract":"This paper proposes a multilayer graph model for community detection based on multiple observations. This scenario is common when different estimators are used to infer graph edges from signals at the nodes, or when various signal measurements are taken. The multilayer network stacks these graph observations at different layers and links replica nodes at adjacent layers. This configuration corresponds to the Cartesian product between the ground truth graph and a path graph, where the number of nodes matches the number of observations. Using the algebraic structure of the Laplacian of the Cartesian multilayer network, we infer a subset of the eigenvectors of the true graph and perform community detection. Experimental results on synthetic graphs demonstrate the accuracy of the method, which outperforms state-of-the-art approaches in correctly detecting graph communities. Finally, we apply our method to distinguish between different brain networks derived from real EEG data collected during motor imagery experiments. We conclude that our approach is promising for identifying graph communities when multiple graph observations are available, and it shows potential for applications such as EEG-based motor imagery applications.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"201-214"},"PeriodicalIF":3.0,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143553482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Penalized Likelihood Approach for Graph Learning in the Presence of Outliers","authors":"Ghania Fatima;Petre Stoica;Prabhu Babu","doi":"10.1109/TSIPN.2025.3540701","DOIUrl":"https://doi.org/10.1109/TSIPN.2025.3540701","url":null,"abstract":"Graph learning is an important problem in the field of graph signal processing. However, the data available in real-world applications are often contaminated with outliers, which makes the application of traditional methods challenging. In this paper, we address this problem by developing an algorithm that effectively learns the graph Laplacian matrix from node signals corrupted by outliers. Specifically, we maximize the penalized log-likelihood of the uncorrupted data, where the penalty is chosen via the false discovery rate (FDR) principle, with respect to both the number of outliers and their locations, as well as the precision matrix of the data under the graph Laplacian constraints. To illustrate the robustness to outliers, we compare our method with two state-of-the-art graph learning methods, one that considers outliers in the data and one that does not, using different performance metrics. Our findings demonstrate that the proposed method efficiently detects the number and positions of outliers and accurately learns the graph in their presence.","PeriodicalId":56268,"journal":{"name":"IEEE Transactions on Signal and Information Processing over Networks","volume":"11 ","pages":"187-200"},"PeriodicalIF":3.0,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143521495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}