From Performance-oriented AI to Production- and Industrial-AI
Tuesday, April 06, 2021
Machine learning systems for complex tasks – such as controlling industrial manufacturing processes in real-time; or writing medical imaging case reports – are becoming increasing sophisticated and consist of a large number of data, model, algorithm, and system elements and modules. Traditional performance-oriented bespoke approaches in the ML community are not suited to meet highly demanding industrial standards beyond performance, such as safety, energy-efficiency, and scalability typically expected in production systems in industries such as healthcare, manufacturing, and transportation.
This talk discusses technical challenges toward production- and industrial-AI from the following aspects: theoretical foundation for panoramic learning with all experiences, compositional strategies for building Pan-ML programs from Lego-like blocks, optimization methods for tunning systems, and systems framework for scaling up and scaling out ML productions. Professor Eric Xing will provide a few examples of our effects to address each of these challenges in the form of first principle formula, new algorithms, software toolkits, and composable systems.
Speaker/s
Prior to joining MBZUAI, Professor Eric Xing was a professor of computer science at Carnegie Mellon University, and the Founder, Chairman, and Chief Scientist of Petuum Inc., which was recognized as a 2018 World Economic Forum Technology Pioneer that builds standardized artificial intelligence development platform and operating system for broad and general industrial AI applications. Additionally, Xing is the Founding Director of the Center for Machine Learning and Health, Carnegie Mellon University, and University of Pittsburgh Medical Center. Professor Xing has also spent time as a Visiting Associate Professor at Stanford University, and as Visiting Research Professor at Facebook Inc.