Yingbo Fan

Title: Innovations of Yingbo Fan in Deep Neural Network Hardware Acceleration

Introduction

Yingbo Fan is an accomplished inventor based in Jiangsu, China. He has made significant contributions to the field of deep learning through his innovative patent. His work focuses on enhancing the efficiency of deep neural network hardware accelerators.

Latest Patents

Yingbo Fan holds a patent for a "Deep Neural Network Hardware Accelerator Based on Power Exponential Quantization." This invention comprises several components, including an AXI-4 bus interface, input and output cache areas, a weighting cache area, a weighting index cache area, an encoding module, a configurable state controller module, and a processing element (PE) array. The design of the input and output cache areas as a line cache structure allows for efficient data handling. The encoder encodes weightings based on an ordered quantization set, which stores the possible values of the absolute weightings after quantization. During operation, the PE unit reads data from the input and weighting index cache areas to perform shift calculations, significantly reducing the need for computing resources, storage resources, and communication bandwidth. This innovation enhances the calculation efficiency of the accelerator.

Career Highlights

Yingbo Fan is affiliated with Southeast University, where he continues to advance research in hardware acceleration for deep learning applications. His work has positioned him as a key figure in the field, contributing to the development of more efficient computational methods.

Collaborations

Yingbo Fan collaborates with notable colleagues, including Shengli Lu and Wei Pang, who contribute to his research endeavors.

Conclusion

Yingbo Fan's innovative work in deep neural network hardware acceleration exemplifies the potential for advancements in computational efficiency. His contributions are paving the way for future developments in the field of artificial intelligence.