Wenwen Ye

Title: Innovations of Wenwen Ye in Neural Network Retraining

Introduction

Wenwen Ye is an accomplished inventor based in Zhejiang, China. She has made significant contributions to the field of neural networks, particularly in the context of memristors. Her innovative approach addresses the challenges posed by aging in hardware systems.

Latest Patents

Wenwen Ye holds a patent for a "Neural network retraining method based on aging sensing of memristors." This method provides a solution for the accuracy rate of hardware online reasoning, which tends to decrease after cross-array aging. By calculating the extreme values of programmable weights under current aging conditions, the method retrains a neural network model to improve accuracy. During retraining, network weights that exceed these extreme values are automatically truncated. Additionally, the method utilizes the sparsity of derivatives to prolong the service life of memristors by preventing unnecessary voltage application.

Career Highlights

Wenwen Ye is affiliated with Zhejiang University, where she continues to advance her research and innovations. Her work has garnered attention for its practical applications in enhancing the performance and longevity of neural network systems.

Collaborations

Wenwen collaborates with notable colleagues, including Xunzhao Yin and Cheng Zhuo. Their combined expertise contributes to the development of cutting-edge technologies in the field of neural networks.

Conclusion

Wenwen Ye's innovative work in neural network retraining showcases her commitment to advancing technology in the realm of memristors. Her contributions are paving the way for more efficient and durable hardware systems.