A5100347934- Neural Networks and Applications
- Robotic Mechanisms and Dynamics
- Robot Manipulation and Learning
- Iterative Learning Control Systems
- Structural Load-Bearing Analysis
- Robotic Path Planning Algorithms
- Adaptive Control of Nonlinear Systems
- Distributed Control Multi-Agent Systems
- Neural Networks Stability and Synchronization
- Fault Detection and Control Systems
- Machine Learning and ELM
- Advanced Memory and Neural Computing
- Model Reduction and Neural Networks
- Structural Behavior of Reinforced Concrete
- Advanced Vision and Imaging
- Modular Robots and Swarm Intelligence
- Neural Networks and Reservoir Computing
- Mechanical stress and fatigue analysis
- Advanced Neural Network Applications
- Piezoelectric Actuators and Control
- Control Systems and Identification
- Structural Engineering and Vibration Analysis
- Vibration and Dynamic Analysis
- Nonlinear Dynamics and Pattern Formation
- Advanced Chemical Sensor Technologies
Lanzhou University
2018-2025
Chinese University of Hong Kong
2024-2025
Harbin Engineering University
2022-2025
Zhoukou Normal University
2018-2025
Beijing University of Chemical Technology
2025
Wuhan Business University
2024
Sun Yat-sen University
2024
Chongqing Institute of Green and Intelligent Technology
2020-2024
Zhangjiakou Academy of Agricultural Sciences
2024
Zhengzhou University
2024
Objective The authors' goal was to develop and evaluate machine-learning-based approaches extracting clinical entities—including medical problems, tests, treatments, as well their asserted status—from hospital discharge summaries written using natural language. This project part of the 2010 Center Informatics for Integrating Biology Bedside/Veterans Affairs (VA) natural-language-processing challenge. Design authors implemented a named entity recognition system text systematically evaluated...
It has been found that there exists an essential similarity between solving equations and controlling dynamic systems: Both errors are expected to decrease zero (or acceptably tiny value) as soon possible. By exploiting such a similarity, researchers have presented investigated continuous-time recurrent neural network models for time-varying problems. To be compatible with digital computers, it is desirable develop discrete-time algorithms from the control perspective performance...
Deep neural networks often suffer from poor performance or even training failure due to the ill-conditioned problem, vanishing/exploding gradient and saddle point problem. In this article, a novel method by acting activation function (GAF) on is proposed handle these challenges. Intuitively, GAF enlarges tiny gradients restricts large gradient. Theoretically, article gives conditions that needs meet and, basis, proves alleviates problems mentioned above. addition, convergence rate of SGD...
Starch is the most abundant organic compound in nature, second only to cellulose. However, conserved structure of native starches limits their properties and applications. Therefore, it often necessary modify them obtain specific properties. As a common green safe physical modification method, pre-gelatinization can improve cold water solubility swelling power starch. The typical preparation methods pre-gelatinized starch include spray drying, extrusion, drum drying. Spray-dried still had...
Enlightened by competitive and collaborative coordination behaviors widely observed in natural swarm systems, this work emphasizes these coordinating modes multirobot systems optimizes system stability along with resource utilization. Then, schemes are constructed to describe model two modes, where a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$k$</tex-math></inline-formula> -winner-take-all concept is...
The diagnosis of brain tumor types generally depends on the clinical experience doctors, and computer-assisted improves accuracy diagnosing types. Therefore, a convolutional neural network based complex networks (CNNBCN) with modified activation function for magnetic resonance imaging classification tumors is presented. structure not manually designed optimized, but generated by randomly graph algorithms. These graphs are mapped into computable generator. CNNBCN model reaches 95.49%, which...
Federated learning (FL) is promising in supporting collaborative applications that involve large datasets, massively distributed data owners and unreliable network connectivity. To protect privacy, existing FL approaches adopt (k,n)-threshold secret sharing schemes, based on the semi-honest assumption for clients, to enable secure multiparty computation local model update exchange which deals with random client dropouts at cost of increasing size. These therefore they are vulnerable...
For the existing repetitive motion generation (RMG) schemes for kinematic control of redundant manipulators, position error always exists and fluctuates. This article gives an answer to this phenomenon presents theoretical analyses reveal that RMG exist a related joint angle error. To remedy weakness solutions, orthogonal projection (OPRMG) scheme is proposed in by introducing method with eliminated theoretically, which decouples space Cartesian constraints considered. The corresponding new...
In fault diagnosis field, inconsistent distribution between training and testing data, resulted from variable working conditions of rotating machinery, inevitably leads to degradation diagnostic performance. To address this issue, study proposes a novel method based on enhanced multi-scale sample entropies balanced adaptation regularization transfer learning. Specifically, different statistics-based are used improve feature discriminability for patterns under each condition enhance...
In this article, a neural dynamics model is constructed and investigated for solving time-dependent Sylvester equation with matrix inversion involved in the process. Besides, to eliminate model, quasi-Newton Broyden-Fletcher-Goldfarb-Shanno method leveraged construct new model. Moreover, global convergence performance effectiveness of two discrete computational models are testified by providing theoretical analyses numerical experiments comparisons existing solutions, respectively. Two...
In this article, three acceleration-level joint-drift-free (ALJDF) schemes for kinematic control of redundant manipulators are proposed and analyzed from perspectives dynamics kinematics with the corresponding tracking error analyses. First, existing ALJDF systematized into a generalized scheme paradox pointing out theoretical existence velocity related to joint drift. Second, remedy deficiency solutions, novel (NALJDF) is decouple Cartesian space theoretically eliminated. Third, in...
Streets functioning as important components of urban public space are not only the primary carriers traffic but also essential spaces for individuals’ daily activities, including recreation and communication. The paper highlights social characteristics streets by integrating them into one single index vitality. application open recourse data combined with empirical research forms foundation a quantitative exploration on street vitality Zhoujiadu Community in Shanghai. Supported ideology...
Precisely controlling the spatial organization of inorganic nanoparticles is essential for many applications, such as sensing, catalytic, and optical/electronic systems. Here, we describe a general robust strategy to fabricate arbitrary quasi-two-dimensional (quasi-2D) nanoparticle assembly structures with tunable size thickness on substrate by using block copolymer brushes templates. To achieve this, patterned amphiphilic were first grown via light-mediated surface-initiated atom transfer...
Aiming at the k-winners-take-all <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(k\text{WTA})$</tex> operation, this paper proposes a gradient-based differential xmlns:xlink="http://www.w3.org/1999/xlink">$k\text{WTA}$</tex> (GD- ) network. After obtaining network, theorems and related proofs are provided to guarantee exponential convergence noise resistance of proposed GD-kWTA Then, numerical simulations conducted substantiate preferable...