EECS Colloquium: Specializing the Computing System for Graph Algorithms — Xuhao Chen, MIT
About the event
Abstract: Numerous AI applications in social networks, e-commerce, biomedicine and security, are driven by graph algorithms. The graph data is massive and sparse, which poses great challenges in computing system design. In this talk, I will explore system design tradeoffs for an important class of graph algorithms — graph mining. I will describe experiences creating abstractions, optimization techniques and automation methodologies for graph mining, across different layers of the system stack, including both software and hardware. As I will demonstrate, cross-layer system specialization can fully unlock the potential of graph computing, and should be used for enabling pervasive AI.
Bio: Xuhao Chen is a Research Scientist at MIT CSAIL, working with Prof. Arvind. Dr. Chen is broadly interested in parallel systems and computer architectures for AI and big-data. His recent work aims to democratize Graph AI by designing efficient algorithms, software and hardware systems. His work has been published in OSDI, ISCA, MICRO, VLDB, ICS, etc. Before joining MIT, Dr. Chen was a Research Fellow at UT Austin.