We are honored to have the following keynote speakers at ICS 2017.
Amir Khosrowshahi
CTO AI Products Group, Co-Founder of Nervana
Amir has an AB and AM from Harvard and PhD from Berkeley. He started his career at Goldman Sachs where he was a derivatives trader. He then worked at startups, returned to school to study neuroscience, and then founded Nervana. He is now VP in the data center group at Intel.
Title:Intel Nervana: Leading the Evolution of Enterprise AI
Abstract: Machine Learning and Deep Learning innovation is enabling the development of a new wave of AI applications and solutions. Intel Nervana’s end to end AI products are optimized for performance, power efficiency, and the scalable development of enterprise AI from data center to edge. Intel Nervana serves as a catalyst for organizations to evolve and transform through deep learning and other computational methods needed for building AI solutions. This session will dive into the Intel Nervana software and hardware stackincluding the Intel Nervana graph, a common execution format for neural networks. Also, this session will explore how Intel Nervana is creating AI hardware that is designed for Deep Learning, featuringunprecedented compute density and true model parallelism.
Long Wang
Senior System Engineer, System Department of Baidu
Long received his computational mathematics PhD from AMSS of the Chinese Academy of Sciences. Before he joined Baidu, he was a professor in Supercomputing Center, CNIC of CAS where he worked on parallel computing algorithms in scientific computing for 10 years. He then worked in Intel’s PRC HPC team , and then he joined Baidu in 2017 for parallel computing in the ML/DL area. Now he is a Senior System Engineer in the System Department of Baidu.
Title: Architecture and Optimization on Baidu’s AI Computing Platform
Abstract: As a leading AI company in China and worldwide, Baidu has very large scale heterogeneous computing platform to speed up training and inference process in different AI applications such as search, NLP, image recognition, etc.
In this talk, we will focus on: 1) architecture of AI platform including heterogeneous accelerators(GPU/FPGA), storage and network topology; 2) our Innovative Technologies and optimization works on this platform to speedup different AI applications; 3) feature and applications of Baidu Cloud computing platform.
