This inventor holds 1 USPTO granted patent and 1 published patent application. Top assignee: Beijing Dajia Internet Information Technology Co., Ltd.. Active years: 2025.
Company Filing History:
Years Active: 2025
Title: Sitong Feng - Innovator in Neural Network Implementation
Introduction
Sitong Feng is a notable inventor based in Hoboken, NJ (US). She has made significant contributions to the field of data processing through her innovative methods and apparatuses. Her work focuses on enhancing the performance and accuracy of neural networks.
Latest Patents
Sitong Feng holds a patent for "Methods and apparatuses for high performance and accuracy fixed-point scale implementation." This patent describes a method to implement a fixed-point scale layer in a neural network for data processing. The method involves receiving fixed-point input data, converting floating-point input data into fixed-point input data, obtaining fixed-point quantization parameters, and mapping the fixed-point scale layer to a fixed-point convolution layer. This innovative approach allows for efficient computation of convolution using matrix multiplication on a GEMM engine.
Career Highlights
Sitong Feng is currently employed at Beijing Dajia Internet Information Technology Co., Ltd. Her work at this company has allowed her to further develop her expertise in neural network technologies and data processing methods.
Collaborations
Sitong collaborates with Ming Kai Hsu, who is also a talented professional in the field. Their partnership enhances the innovative capabilities of their projects and contributes to the advancement of technology.
Conclusion
Sitong Feng's contributions to neural network implementation demonstrate her commitment to innovation in data processing. Her patent and career achievements reflect her expertise and dedication to advancing technology in this field.