On the opportunities of the design for the next generation of collective communication libraries for AI-centric workflows
Binhang Yuan
Hong Kong University of Science and Technology (HKUST)
Talk
Binhang Yuan is an Assistant Professor at the Department of Computer Science and Engineering (CSE), the Hong Kong University of Science and Technology (HKUST). He received his Ph.D. and master's degrees from Rice University and his bachelor's degree from Fudan University. Before joining HKUST, he was a Postdoc at the Swiss Federal Institute of Technology Zurich (ETH Zurich). His main research interests are in data management systems for machine learning, distributed and decentralized machine learning systems.
AG 1, AG 2, AG 3, INET, AG 4, AG 5, D6, SWS, RG1, MMCI
The design and implementation of the next generation of collective communication libraries present significant opportunities for enhancing the end-to-end performance of AI-centric workflows, including heterogeneous compute unit support, deep Integration with AI frameworks, and fault tolerance and Reliability. The development of collective communication libraries that address these areas will enhance the performance of AI applications and broaden their potential use cases, supporting more complex, dynamic, and scalable AI workflows.
Contact
Iris Wagner
+49 681 9325 3500
--email hidden
passcode not visible
Iris Wagner, 07/15/2024 14:02 -- Created document.