Company Filing History:
Years Active: 2023-2025
Title: Innovations of Inventor Kai-Jiun Yang
Introduction
Kai-Jiun Yang is a notable inventor based in Zhubei, Taiwan. He has made significant contributions to the fields of data compression and deep learning acceleration. With a total of 2 patents, his work showcases innovative methods that enhance technology efficiency.
Latest Patents
One of his latest patents is a data compression method, data compression system, and operation method of a deep learning acceleration chip. This invention involves obtaining a filter coefficient tensor matrix of a deep learning model and performing a matrix decomposition procedure. The result is a sparse tensor matrix and a transformation matrix, which is orthonormal. The sparse tensor matrix is then compressed and stored in memory, allowing for efficient convolution operations by the deep learning acceleration chip.
Another significant patent by Yang is a MIMO signal symbol detection and search method, along with a decoding circuit and receiving antenna system. This method includes obtaining a symbol search tree and sorting candidate symbols at each layer. By traversing these symbols and updating thresholds based on cumulative partial Euclidean distances, the method efficiently outputs an estimated symbol combination.
Career Highlights
Kai-Jiun Yang is associated with the Industrial Technology Research Institute, where he continues to innovate and develop new technologies. His work has implications for various applications, particularly in enhancing data processing and communication systems.
Collaborations
Yang collaborates with esteemed colleagues such as Chi-Tien Sun and Shang-Ho Tsai, contributing to a dynamic research environment that fosters innovation.
Conclusion
Kai-Jiun Yang's contributions to data compression and deep learning acceleration exemplify the impact of innovative thinking in technology. His patents reflect a commitment to advancing the field and improving efficiency in data processing.