Authors: Wei Li, Chandan Damannagari, Scott Lenker
AI applications are starting to appear in almost all aspects of our everyday lives, from healthcare and finance to entertainment and environmental protection. But a large number of AI applications never make it from concept to implementation, and an even larger amount never even get started. How do we enable more data scientists and developers to quickly create the path from data to insights with the data and compute resources available to them? The key to traversing this path and enabling AI everywhere is software that both unlocks hardware performance and maximizes developer productivity.
In this keynote presentation from the oneAPI DevSummit for AI 2022, Wei Li, VP/GM AI & Analytics at Intel, outlines the Intel AI software strategy, including:
The core pillars of oneAPI and Intel’s AI software strategy
How Intel software libraries help deliver high-performance AI applications
Intel’s collaboration with Google to include oneDNN AI optimizations as default in TensorFlow 2.9
The launch of industry use case AI Reference Kits, together with Jean-Luc Chatelain, CTO of Accenture Applied Intelligence
The collaboration between Intel and Huggingfaceon the Optimum open source library to deliver hardware-optimizations for Transformer models.
About our expert
Dr. Wei Li is the Vice President and General Manager of AI & Analytics at Intel. After starting his career as a computer scientist for supercomputers, he received his Ph.D. in Computer Science from Cornell University and taught Advanced Compiling Techniques at Stanford University. With a passion for technology, strategy, and execution, he has been instrumental in Intel's multi-billion-dollar AI revenue growth. You can find him on LinkedIn.
