A5102927906- Tensor decomposition and applications
- Advanced Neural Network Applications
- CCD and CMOS Imaging Sensors
- Generative Adversarial Networks and Image Synthesis
- Energy Load and Power Forecasting
- Higher Education Governance and Development
- Advanced Image and Video Retrieval Techniques
- Advanced Computing and Algorithms
- Computational Physics and Python Applications
- Ferroelectric and Negative Capacitance Devices
- Vehicle License Plate Recognition
- Manufacturing Process and Optimization
- Distributed and Parallel Computing Systems
- Knowledge Management and Sharing
- Embedded Systems Design Techniques
- University-Industry-Government Innovation Models
- Advanced Memory and Neural Computing
- Semiconductor materials and devices
Harbin Engineering University
2024
California State University, Fullerton
2021-2023
MicroPort (China)
2022
Nowadays, from companies to academics, researchers across the world are interested in developing Deep Neural Networks (DNNs) due their incredible feats various applications, such as image recognition, playing complex games, and large-scale information retrieval web search. However, when people enjoy advantages of DNNs, high computational power demands on resource-constrained electronic devices DNN model have received more attention. Optimizing model, compression, is crucial ensure wide...
The fast development of object detection techniques has attracted attention to developing efficient Deep Neural Networks (DNNs). However, the current state-of-the-art DNN models can not provide a balanced solution among accuracy, speed, and model size. This paper proposes an real-time framework on resource-constricted hardware devices through software co-design. Tensor Train (TT) decomposition is proposed for compressing YOLOv5 model. By unitizing unique characteristics given by TT...
In this article, a neural spin-orbit torque (NSOT)-based emerging device technology reconfigurable fabric is developed and assessed as low-leakage power alternative to the use of static random access memory (SRAM)-based lookup table (LUT). While other recent designs have explored utilizing spintronic devices digital cell replace SRAM-based configuration in field programmable gate arrays (FPGAs), NSOT-LUT exploits small, modularized, validated artificial networks (ANNs) together with behavior...
This paper aims to explore the issues of human resource management reform in universities and provide some preliminary exploration suggestions. As important institutions for talent cultivation, face new challenges changes management, requiring continuous adaptation improvement. first introduces background current situation universities, then discusses necessity reform. Through empirical research case analysis, it summarizes key elements successful experiences universities. Finally, proposes...
The fast development of object detection techniques has attracted attention to developing efficient Deep Neural Networks (DNNs). However, the current state-of-the-art DNN models can not provide a balanced solution among accuracy, speed, and model size. This paper proposes an real-time framework on resource-constrained hardware devices through software co-design. Tensor Train (TT) decomposition is proposed for compressing YOLOv5 model. By unitizing unique characteristics given by TT...
Nowadays, from companies to academics, researchers across the world are interested in developing recurrent neural networks due their incredible feats various applications, such as speech recognition, video detection, predictions, and machine translation. However, advantages of accompanied by high computational power demands, which a major design constraint for electronic devices with limited resources used network implementations. Optimizing networks, model compression, is crucial ensure...
Deep learning methods have exhibited the great capacity to process object detection tasks, offering a practical and viable approach in many applications. When researchers advanced deep models improve their performance, model derived from algorithmic improvement may itself require complementary increases computational power demands. Recently, compression pruning techniques received more attention promote wide employment of DNN model. Although these achieved remarkable class imbalance issue...
In order to solve the problem of communication interference and distance caused by rapid pacing system when establishing atrial fibrillation model, a low-power implantable based on 433 MHz frequency form star network is designed. The includes an pacemaker, programmer head, software. pacemaker composed wireless module, ECG monitoring power management module. head acts as intermediate node in controlled PC software program each pacemaker. This article introduces hardware design flow part...