{"title":"Efficient Linear Discriminant Analysis Based on Randomized Low-Rank Approaches","authors":"Yujie Wang, Weiwei Xu, Lei Zhu","doi":"10.1109/tnnls.2025.3547013","DOIUrl":"https://doi.org/10.1109/tnnls.2025.3547013","url":null,"abstract":"","PeriodicalId":13303,"journal":{"name":"IEEE transactions on neural networks and learning systems","volume":"12 1","pages":""},"PeriodicalIF":10.4,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143672442","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}
Kenji Kashima, Ryota Yoshiuchi, Ran Wang, Yu Kawano
{"title":"A Unified Framework for Dynamics Modeling and Control Design Using Deep Learning With Side Information on Stabilizability.","authors":"Kenji Kashima, Ryota Yoshiuchi, Ran Wang, Yu Kawano","doi":"10.1109/TNNLS.2025.3543926","DOIUrl":"10.1109/TNNLS.2025.3543926","url":null,"abstract":"<p><p>This article presents a unified framework for dynamics modeling and control design using deep learning, focusing on incorporating prior side information on stabilizability. Control theory provides systematic techniques for designing feedback systems while ensuring fundamental properties such as stabilizability, which are crucial for practical control applications. However, conventional data-driven approaches often overlook or struggle to explicitly incorporate such control properties into learned models. To address this, we introduce a novel neural network (NN)-based approach that concurrently learns the system dynamics, a stabilizing feedback controller, and a Lyapunov function for the closed-loop system, thus explicitly guaranteeing stabilizability in the learned model. Our proposed deep learning framework is versatile and applicable across a wide range of control problems, including safety control, -gain control, passivation, and solutions to Hamilton-Jacobi inequalities. By embedding stabilizability as a core property within the learning process, our method allows for the development of learned models that are not only data-driven but also grounded in control-theoretic guarantees, greatly enhancing their utility in real-world control applications. This article includes examples that demonstrate the effectiveness of this approach, showcasing the stability and control performance improvements achieved in various control scenarios. The methods proposed in this article can be easily applied to modeling without control design. The code has been open-sourced and is available at https://github.com/kashctrl/Deep_Stabilizable_Models.</p>","PeriodicalId":13303,"journal":{"name":"IEEE transactions on neural networks and learning systems","volume":"PP ","pages":""},"PeriodicalIF":10.2,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143669796","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}
{"title":"Diffusion Model-Based Path Follower for a Salamander-Like Robot.","authors":"Zhiang Liu, Yang Liu, Yongchun Fang","doi":"10.1109/TNNLS.2025.3549307","DOIUrl":"10.1109/TNNLS.2025.3549307","url":null,"abstract":"<p><p>Salamander-like robots, renowned for their versatile locomotion, present unique challenges in the development of effective path-following controllers due to their distinctive movement patterns and complex body structures. Conventional path-following controllers, while effective for various bionic robots, struggle with the intricate modeling for salamander-like robots and often require laborious manual tuning. Conversely, learning-based methods offer promising alternatives but face issues such as reliance on environmental interactions, short-sighted prediction, and irrational design of state space and reward function. To overcome these limitations, this article proposes a diffusion model-based hierarchical control framework that treats path tracking as a sequence generation problem. The diffusion model's capability to model joint distributions of state, action, and reward sequences enables it to outperform other learning-based approaches in efficient data utilization, stable training, and long-horizon dependency modeling. Our framework integrates a high-level policy driven by guided diffusion with a low-level controller for parsing commands into executable movements via inverse kinematics, reducing the action space and improving learning efficiency. In addition, we design a more reasonable state space and reward function tailored to the path-following task, addressing shortcomings in prior learning-based controllers. Furthermore, we optimize the diffusion model (DM) by developing lightweight network architectures and incorporating advanced attention mechanisms, to ensure its practical deployment on physical robots with limited computational resources, without compromising performance. Extensive simulations and real-world experiments demonstrate the framework's effectiveness, efficiency, and robustness in diverse path-following tasks for salamander-like robots, marking a significant advancement in the control of biomimetic robots.</p>","PeriodicalId":13303,"journal":{"name":"IEEE transactions on neural networks and learning systems","volume":"PP ","pages":""},"PeriodicalIF":10.2,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143669798","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}
Yue Zhao, Maoguo Gong, Mingyang Zhang, A K Qin, Fenlong Jiang, Jianzhao Li
{"title":"SPCNet: Deep Self-Paced Curriculum Network Incorporated With Inductive Bias.","authors":"Yue Zhao, Maoguo Gong, Mingyang Zhang, A K Qin, Fenlong Jiang, Jianzhao Li","doi":"10.1109/TNNLS.2025.3544724","DOIUrl":"10.1109/TNNLS.2025.3544724","url":null,"abstract":"<p><p>The vulnerability to poor local optimum and the memorization of noise data limit the generalizability and reliability of massively parameterized convolutional neural networks (CNNs) on complex real-world data. Self-paced curriculum learning (SPCL), which models the easy-to-hard learning progression from human beings, is considered as a potential savior. In spite of the fact that numerous SPCL solutions have been explored, it still confronts two main challenges exactly in solving deep networks. By virtue of various designed regularizers, existing weighting schemes independent of the learning objective heavily rely on the prior knowledge. In addition, alternative optimization strategy (AOS) enables the tedious iterative training procedure, thus there is still not an efficient framework that integrates the SPCL paradigm well with networks. This article delivers a novel insight that attention mechanism allows for adaptive enhancement in the contribution of diverse instance information to the gradient propagation. Accordingly, we propose a general-purpose deep SPCL paradigm that incorporates the preferences of implicit regularizer for different samples into the network structure with inductive bias, which in turn is formalized as the self-paced curriculum network (SPCNet). Our proposal allows simultaneous online difficulty estimation, adaptive sample selection, and model updating in an end-to-end manner, which significantly facilitates the collaboration of SPCL to deep networks. Experiments on image classification and scene classification tasks demonstrate that our approach surpasses the state-of-the-art schemes and obtains superior performance.</p>","PeriodicalId":13303,"journal":{"name":"IEEE transactions on neural networks and learning systems","volume":"PP ","pages":""},"PeriodicalIF":10.2,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143669800","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}
Boyuan Yang, Jinyuan Zhang, Ruonan Liu, Di Lin, Ping Li, C. L. Philip Chen
{"title":"Point-to-Set Metric-Gated Mixture of Experts for Multisource Domain Adaptation Fault Diagnosis","authors":"Boyuan Yang, Jinyuan Zhang, Ruonan Liu, Di Lin, Ping Li, C. L. Philip Chen","doi":"10.1109/tnnls.2025.3548894","DOIUrl":"https://doi.org/10.1109/tnnls.2025.3548894","url":null,"abstract":"","PeriodicalId":13303,"journal":{"name":"IEEE transactions on neural networks and learning systems","volume":"37 1","pages":""},"PeriodicalIF":10.4,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143661450","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}
Xi Yang, Wenjiao Dong, De Cheng, Nannan Wang, Xinbo Gao
{"title":"TIENet: A Tri-Interaction Enhancement Network for Multimodal Person Reidentification","authors":"Xi Yang, Wenjiao Dong, De Cheng, Nannan Wang, Xinbo Gao","doi":"10.1109/tnnls.2025.3544679","DOIUrl":"https://doi.org/10.1109/tnnls.2025.3544679","url":null,"abstract":"","PeriodicalId":13303,"journal":{"name":"IEEE transactions on neural networks and learning systems","volume":"59 1","pages":""},"PeriodicalIF":10.4,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143661452","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}