Wenqian Pan

Title: Innovations of Wenqian Pan in Convolutional Neural Networks

Introduction

Wenqian Pan is a notable inventor based in Hubei, China. He has made significant contributions to the field of convolutional neural networks, particularly in the development of on-chip learning systems. With a total of 2 patents, his work is paving the way for advancements in artificial intelligence and memory-based systems.

Latest Patents

Wenqian Pan's latest patents include a "Convolutional Neural Network On-Chip Learning System Based on Non-Volatile Memory." This innovative system comprises an input module, a convolutional neural network module, an output module, and a weight update module. The on-chip learning of the convolutional neural network module utilizes the characteristics of memristors to implement synaptic functions. The system effectively converts input signals into voltage signals and processes them layer-by-layer to produce network outputs. The weight update module adjusts the conductance values of the memristors based on the output results, allowing for dynamic updates of the convolutional kernel values or synaptic weights.

Another significant patent is the "Memory-Based Convolutional Neural Network System." This system includes various modules such as an input module, convolution layer circuit module, pooling layer circuit module, activation function module, fully connected layer circuit module, softmax function module, and an output module. In this design, convolution kernel values or synapse weights are stored in NOR FLASH units. The system processes input signals into voltage signals, convolves them with kernel values, and performs classification through a series of operations, ultimately normalizing the results into probability values.

Career Highlights

Wenqian Pan is affiliated with Huazhong University of Science and Technology, where he continues to engage in groundbreaking research. His work focuses on enhancing the efficiency and effectiveness of convolutional neural networks, contributing to the broader field of artificial intelligence.

Collaborations

Wenqian Pan collaborates with esteemed colleagues, including Xiangshui Miao and Yi Bu Li, who contribute to his research endeavors. Their combined expertise fosters an environment of innovation and discovery.

Conclusion

Wenqian Pan's contributions to convolutional neural networks exemplify the potential of innovative technologies in artificial intelligence. His patents reflect a commitment to advancing the field and improving the capabilities of on-chip learning systems.