Co-sponsored by Signal Processing Society (SPS) and Computer Society (CS)

Title: Khronos Standards for Neural Network Acceleration and Deployment

Abstract: Khronos Group is a not for profit, member-funded consortium to create royalty-free open standards for hardware acceleration. OpenVX is an API for computer vision and neural network acceleration, especially important in real-time and safety-critical use cases. Khronos Group is also readying NNEF standard interchange format to transfer networks trained in deep learning frameworks to optimized inference engines.This talk gives an overview of Khronos standards related to neural networks and computer vision. A set of examples for neural networks and computer vision mapped to graph API will also be discussed.

Speaker Bio:

Radhakrishna (Radha) Giduthuri is a software architect at Advanced Micro Devices (AMD) focusing on development of computer vision and neural network acceleration libraries for AMD GPUs. He has extensive background with software design and performance tuning for various computer architectures ranging from General Purpose DSPs, Customizable DSPs, Media Processors, Heterogeneous Processors, GPUs, and several CPUs. He is a member of Khronos OpenVX and NNEF working groups, and editor of OpenVX safety-critical specification. For several years, he was a member of SMPTE video compression standardizing committee. He is an active member of IEEE Signal Processing Society and winner of outstanding leadership and professional services award for IEEE Central Area in 2016. Radhakrishna holds M.Tech. degree from IIT Kharagpur, India.