Company Filing History:
Years Active: 2025
Title: Innovations of Wenjie Dai in Neural Network Acceleration
Introduction
Wenjie Dai is a prominent inventor based in Wuhan, China. He has made significant contributions to the field of neural network technology, particularly through his innovative patents. His work focuses on enhancing the efficiency and performance of neural network inference processes.
Latest Patents
Wenjie Dai holds 2 patents, one of which is an anti-radiation low-latency chip designed for neural network inference acceleration. This invention proposes a novel approach to deploying a neural network inference acceleration chip at the frontend of a detector. The chip utilizes a streaming architecture that aligns each pipeline stage with major neural network layers. It effectively balances limited on-chip memory resources while supporting large-sized inputs. The design incorporates layer parallelism, channel parallelism, and convolution kernel parallelism to optimize performance. The application aims to improve the intelligence level of future detector hardware by implementing fine-grained streaming architecture, flexible compression, quantization, and anti-radiation digital chip design technologies. This results in high throughput and low on-chip memory consumption while maintaining anti-radiation and low-latency inference capabilities.
Career Highlights
Wenjie Dai is affiliated with Central China Normal University, where he continues to advance his research and development in neural network technologies. His academic background and innovative mindset have positioned him as a key figure in the field.
Collaborations
Wenjie has collaborated with notable colleagues, including Le Xiao and Guoxiang Zhang, who contribute to his research endeavors and enhance the impact of his inventions.
Conclusion
Wenjie Dai's contributions to neural network technology through his innovative patents demonstrate his commitment to advancing the field. His work not only addresses current technological challenges but also paves the way for future developments in intelligent hardware.