IEEE Transactions on Cybernetics最新文献_第2页

Causal Semantic Alignment for LLM-Based Time Series Forecasting. 基于llm的时间序列预测的因果语义对齐。

IF 11.3 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-10-09 DOI: 10.1109/TCYB.2026.3737567

Kexuan Zhang, Xiaobei Zou, Cesare Alippi, Gary G Yen, Yang Tang

{"title":"Causal Semantic Alignment for LLM-Based Time Series Forecasting.","authors":"Kexuan Zhang, Xiaobei Zou, Cesare Alippi, Gary G Yen, Yang Tang","doi":"10.1109/TCYB.2026.3737567","DOIUrl":"10.1109/TCYB.2026.3737567","url":null,"abstract":"Recent advances in large language models (LLMs) have opened new possibilities for time series forecasting by enabling alignment between temporal patterns and pretrained word embeddings. However, most LLM-based forecasting methods overlook the heterogeneous nature of time series, where stable semantics and dynamic variations are often entangled. Such entanglement may introduce environmentally induced variations during cross-modal alignment, reducing the stability of semantic representations. To address this issue, we propose a causal variable-level alignment Transformer (CVAformer), a variable-level semantic alignment framework for time series forecasting. CVAformer explicitly decomposes each variable into invariant and dynamic representations before alignment and performs dynamic-factor conditioning to improve alignment robustness. In addition, CVAformer employs NCBlocks to model cross-variable interactions that support semantic alignment and forecasting. Extensive experiments across long-term, short-term, few-shot, and zero-shot forecasting settings demonstrate that CVAformer consistently achieves competitive or state-of-the-art performance on diverse benchmarks. Further analyses verify the effectiveness of variable-level semantic alignment and invariant-dynamic decomposition, providing a promising direction for integrating representation learning and causal reasoning in LLM-based time series forecasting.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"PP ","pages":""},"PeriodicalIF":11.3,"publicationDate":"2026-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149147759","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optical Transport Network Design via Distributionally Robust Nonlinear Optimization 基于分布鲁棒非线性优化的光传输网络设计

IF 11.8 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-10-05 DOI: 10.1109/tcyb.2026.3737718

Fanghao Yin, Xiang Li, Qiang Shen

引用次数: 0

Bidirectional Condition-Guided Denoising Diffusion Model for Industrial Anomaly Detection 工业异常检测的双向条件引导去噪扩散模型

IF 11.8 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-10-05 DOI: 10.1109/tcyb.2026.3738607

Liwei Wang, Jinhai Liu, Huaguang Zhang, Yuan Zhong, Zhen Zhao

引用次数: 0

An Optimization Framework for Multirobot Space Exploration With Hybrid RRT-FE and Market Mechanisms 基于RRT-FE和市场机制的多机器人空间探索优化框架

IF 11.8 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-10-01 DOI: 10.1109/tcyb.2026.3736817

Lin Zhang, Zhongliang Yu, Xiaojie Su, Huafeng Qin, Rene-Vinicio Sanchez Loja

引用次数: 0

Multiview Interaction-Aware Defense Against Composite Attack in Federated Learning 联邦学习中多视图交互感知防御复合攻击

IF 11.8 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-09-29 DOI: 10.1109/tcyb.2026.3735876

Mahmudul Hasan, Md Palash Uddin, Yong Xiang, John Yearwood, Yushu Zhang, Longxiang Gao

引用次数: 0

Multilevel Fault Information Fusion Based on Deep Block-Wise Semi-Nonnegative Matrix Factorization for Industrial Fault Detection and Diagnosis. 基于深度分块半非负矩阵分解的多级故障信息融合在工业故障检测与诊断中的应用。

IF 11.3 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-09-25 DOI: 10.1109/TCYB.2026.3732779

Xiangyin Kong, Bingxin Zhang, Sida Chai, Xiaoyu Jiang

引用次数: 0

Multiobjective Evolutionary Diffusion Transformer Acceleration. 多目标演化扩散变压器加速。

IF 11.3 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-09-25 DOI: 10.1109/TCYB.2026.3731615

Yao Zhou, Zhang Yi, Gary G Yen

{"title":"Multiobjective Evolutionary Diffusion Transformer Acceleration.","authors":"Yao Zhou, Zhang Yi, Gary G Yen","doi":"10.1109/TCYB.2026.3731615","DOIUrl":"10.1109/TCYB.2026.3731615","url":null,"abstract":"Diffusion Transformers (DiTs) have demonstrated outstanding performance in the field of image synthesis. However, the prohibitive computational overhead of model inference has been a major hurdle for deployment in low-latency scenarios. Based on the repetitive nature of the denoising process, we propose a multiobjective evolutionary algorithm for DiT acceleration, dubbed MEDA. Particularly, the conflict between the computational cost of the sampling process and the degradation in generation quality is explicitly modeled as a two-objective optimization problem, and similar intermediate features in adjacent sampling steps are stored and reused according to evolutionary-optimized schedules. Besides, a dynamic condition evaluation strategy is designed and incorporated into the evolutionary algorithm to maintain the generation performance of accelerated DiT models across different condition inputs. In addition, a latent objective normalization strategy is introduced to provide a bounded objective space for the evolution process. Experiments across various DiT architectures demonstrate that the proposed MEDA can significantly accelerate the sampling process while maintaining or even further enhancing generation quality relative to the original baseline models.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"PP ","pages":""},"PeriodicalIF":11.3,"publicationDate":"2026-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149060566","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dynamic Event-Triggered Adaptive Fault-Tolerant Bipartite Formation-Containment Control for MASs via UIO. 基于UIO的动态事件触发自适应容错二部队形控制。

IF 11.3 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-09-24 DOI: 10.1109/TCYB.2026.3725590

Haichuan Xu, Fanglai Zhu, Xufeng Ling

{"title":"Dynamic Event-Triggered Adaptive Fault-Tolerant Bipartite Formation-Containment Control for MASs via UIO.","authors":"Haichuan Xu, Fanglai Zhu, Xufeng Ling","doi":"10.1109/TCYB.2026.3725590","DOIUrl":"https://doi.org/10.1109/TCYB.2026.3725590","url":null,"abstract":"This article investigates the adaptive fault-tolerant time-varying bipartite formation-containment (TVBFC) problem of multiagent systems (MASs) subject to sensor and actuator faults. The MASs consist of multiple formation leaders (FLs), followers, and one autonomous tracking leader (TL). To begin with, for the followers and FLs to deal with the sensor faults, an output transformation is introduced, and thus an auxiliary output that decouples sensor faults is obtained. Subsequently, for coping with the nonlinear disturbance and actuator faults, a new control-augmented unknown input (CAUI) is constructed. Then, an unknown input observer (UIO), which is able to estimate the states, the CAUI, and the sensor faults simultaneously, is designed via an interval observer. Besides, a Luenberger observer (LO) is designed for estimating the state and sensor fault of the autonomous TL. An adaptive UIO-based dynamic event-triggered control (DETC) protocol is proposed for realizing TVBFC of MASs, and the Zeno behaviors of FLs and followers are also excluded. Finally, three simulation examples are given to demonstrate the effectiveness and advantages of the proposed methods.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"PP ","pages":""},"PeriodicalIF":11.3,"publicationDate":"2026-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149038328","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Output Feedback Cooperative-Antagonistic Average Consensus in High-Order Multiagent Systems Over Directed Graphs. 有向图上高阶多智能体系统的输出反馈合作-对抗平均一致性。

IF 11.3 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-09-23 DOI: 10.1109/TCYB.2026.3727703

Xiaohua Ge, Dario Giuseppe Lui, Alberto Petrillo, Stefania Santini

{"title":"Output Feedback Cooperative-Antagonistic Average Consensus in High-Order Multiagent Systems Over Directed Graphs.","authors":"Xiaohua Ge, Dario Giuseppe Lui, Alberto Petrillo, Stefania Santini","doi":"10.1109/TCYB.2026.3727703","DOIUrl":"10.1109/TCYB.2026.3727703","url":null,"abstract":"Output average consensus of cooperative/ cooperative-antagonistic multiagent systems (MASs) is of great significance in the engineering field, but is rarely addressed for general linear high-order networks interacting over directed graphs. This problem is herein tackled and solved via a novel hierarchical distributed output feedback control architecture, where the higher level embeds a distributed reference generator while the lower one is characterized by a proportional-integral-derivative (PID)-like control structure. The former is responsible for gathering the networked information to compute the output average agreement, which is exploited by the latter for the achievement of the coordination task. Combining the regulator equations and the static output feedback (SOF) procedure, properly extended in the distributed context, a well-defined closed-loop network is derived, and the control gain tuning procedure is provided. The versatility of the proposed approach in also handling the heterogeneous case, arising when agents do not share the same dynamical model, is also proved. Theoretical results establish the stability conditions via the Lyapunov-based analysis and the input-to-state stability (ISS) theory, and prove the accomplishment of the output unsigned/signed average consensus objectives with transient and steady-state performance improvements. Illustrative examples further corroborate the validity of the theoretical results.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"PP ","pages":""},"PeriodicalIF":11.3,"publicationDate":"2026-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149030159","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A Distributed Protocol to Virtualize All-to-All Connectivity in a Class of Multiagent Systems. 一类多智能体系统中虚拟化全对全连接的分布式协议。

IF 11.3 1区计算机科学

IEEE Transactions on Cybernetics Pub Date : 2026-09-23 DOI: 10.1109/TCYB.2026.3732790

Luigi D'Alfonso, Giuseppe Fedele

{"title":"A Distributed Protocol to Virtualize All-to-All Connectivity in a Class of Multiagent Systems.","authors":"Luigi D'Alfonso, Giuseppe Fedele","doi":"10.1109/TCYB.2026.3732790","DOIUrl":"10.1109/TCYB.2026.3732790","url":null,"abstract":"This article addresses the challenge of achieving virtual full connectivity in multiagent systems (MASs) through a novel distributed protocol. Leveraging finite-time convergent observers, the proposed method enables each agent to estimate the states of all nonneighboring agents, facilitating seamless coordination in scenarios where direct communication links are absent. Unlike conventional approaches that focus on leader-following or partial connectivity frameworks, this method ensures mutual state estimation among all agents, thus enabling scalable control. The proposed model incorporates physical constraints, such as bounded agent velocity, which ensures practical realizability and allows the appropriate tuning of observer parameters to guarantee convergence within a finite time. A smooth transition function is introduced to integrate estimated states into the interaction dynamics, ensuring stability and continuity during the transition to a virtually fully connected topology. The system's properties, including collision avoidance, are preserved throughout this transition by deriving a new potential function that accounts for both directly connected and virtually connected agents. The theoretical contributions are validated through a rigorous analysis of convergence properties and the preservation of key interaction rules. Numerical results demonstrate the applicability of the proposed approach.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"PP ","pages":""},"PeriodicalIF":11.3,"publicationDate":"2026-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149030561","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0