Company Filing History:
Years Active: 2022-2024
Title: Inventor Spotlight: Stephen Sangho Youn
Introduction
Stephen Sangho Youn is an innovative inventor based in Bellevue, Washington, with a remarkable portfolio of six patents to his name. His work focuses on advancements in deep neural network technologies and associated architectures, paving the way for enhanced efficiencies in computation and memory management.
Latest Patents
Among Youn's recent contributions to the field are two notable patents. The first is titled "Semi-programmable and reconfigurable co-accelerator for a deep neural network with normalization or non-linearity." This invention details devices that utilize a configurable stacked architecture to implement a fixed function datapath for accelerating specific operations within deep neural networks. The architecture includes configurable micro-execution units designed to handle a series of vector, scalar, reduction, broadcasting, and normalization processes for DNN layer operations, providing a customizable solution tailored to the specific neural network needs.
His second recent patent is "Banked memory architecture for multiple parallel datapath channels in an accelerator," which offers devices and methods for employing a banked memory structure in accelerators. This innovation segments and isolates data flows within the datapath and memory of the accelerator, enabling each data channel to operate with its own dedicated register memory bank. The method also incorporates a memory address decoder to accurately allocate local variables to appropriate memory banks.
Career Highlights
Stephen Sangho Youn is currently associated with Microsoft Technology Licensing, LLC, where he continues to contribute significantly to technological advancements in the realm of deep learning and accelerator architectures. His role at Microsoft has allowed him to work at the forefront of innovation, establishing himself as a key player in the tech industry.
Collaborations
Youn collaborates with a talented group of professionals, including coworkers Steven Karl Reinhardt and Hui Geng. Together, they work synergistically to push the boundaries of technology, enhance computational efficiencies, and develop impactful solutions for modern challenges in computing and artificial intelligence.
Conclusion
In conclusion, Stephen Sangho Youn's contributions through his patents and collaborative efforts have made a significant impact on the field of deep learning and high-performance computing. His innovative spirit and dedication to advancing technology play a vital role in shaping the future of neural networks and data processing architectures.