Jing-Ren Chen

Title: Innovations of Jing-Ren Chen in Neural Network Quantization

Introduction

Jing-Ren Chen is a notable inventor based in Changhua County, Taiwan. He has made significant contributions to the field of neural networks, particularly in the area of data quantization. His innovative approach aims to enhance the efficiency of neural network models.

Latest Patents

Jing-Ren Chen holds a patent for a "Separate quantization method of forming a combination of 4-bit and 8-bit data of a neural network." This patent discloses a method that utilizes a calibration approach to determine activation thresholds for various layers of a neural network model. The method allows for the quantization of activations into 8-bit formats while allocating weights into 4-bit and 8-bit categories based on a predetermined ratio. This innovation effectively reduces the size of the neural network model while maintaining performance.

Career Highlights

Jing-Ren Chen is affiliated with National Yang Ming Chiao Tung University, where he continues to advance research in neural networks and data quantization. His work is pivotal in the ongoing development of more efficient neural network architectures.

Collaborations

Jing-Ren Chen collaborates with esteemed colleagues such as Tien-Fu Chen and Chien-Chih Chen. Their combined expertise contributes to the advancement of innovative solutions in the field of artificial intelligence.

Conclusion

Jing-Ren Chen's contributions to neural network quantization represent a significant step forward in the field of artificial intelligence. His innovative methods are paving the way for more efficient and effective neural network models.