Feng Yuan

Title: Innovations by Feng Yuan in Neural Network Optimization

Introduction

Feng Yuan is an accomplished inventor based in Shanghai, China. He has made significant contributions to the field of deep learning and neural networks. His innovative work focuses on optimizing low precision inference models, which are crucial for the deployment of deep neural networks.

Latest Patents

Feng Yuan holds a patent for "Optimizing low precision inference models for deployment of deep neural networks." This patent describes systems, apparatuses, and methods that provide technology for optimizing an inference neural network model that performs asymmetric quantization. The technology involves generating a quantized neural network, where model weights are quantized as signed integer values. Additionally, the input layer of the neural network is configured to quantize input values as unsigned integer values. The process includes generating a weights accumulation table based on the quantized model weights and a kernel size for the neural network, as well as creating an output restoration function for the output layer based on these elements. The technology also encompasses per-input channel quantization and mixed-precision auto-tuning. Feng Yuan has 1 patent to his name.

Career Highlights

Feng Yuan is currently employed at Intel Corporation, where he continues to advance his research and development in neural network technologies. His work at Intel has positioned him as a key player in the field of artificial intelligence and machine learning.

Collaborations

Feng Yuan collaborates with notable colleagues, including Jiong Gong and Yong Wu. Their combined expertise contributes to the innovative projects at Intel Corporation.

Conclusion

Feng Yuan's contributions to the optimization of neural networks demonstrate his commitment to advancing technology in artificial intelligence. His patent and work at Intel Corporation highlight his role as a leading inventor in this rapidly evolving field.