Speaker: Vijay Janapa Reddi
Abstract: Deep Learning is transforming the field of machine learning (ML) from theory to practice. It has also sparked a renaissance in computer system design. Both academics and the industry are scrambling to integrate ML-centric solutions into their products. Despite the breakneck pace of innovation, there is a crucial issue affecting the research and industry communities at large: how to enable fair and useful benchmarking of ML software frameworks, ML hardware accelerators, and ML systems. The ML field stands in need of systematic benchmarking that is both representative of real-world use-cases and useful for making neutral comparisons across different software and hardware platforms. MLPerf answers the call. MLPerf is a machine learning benchmark standard, and suite, driven by academia and industry (50+ companies). The talk describes the design principles behind MLPerf. It discusses the challenges and opportunities in developing a benchmark for the industry that tackles the complexity, heterogeneity, and scale of ML training and inference systems. See [ Ссылка ] for details.
Bio: Vijay Janapa Reddi is an associate professor in the John A. Paulson School of Engineering and Applied Sciences (SEAS) at Harvard University. He is also the inference chair of MLPerf - A broad community focused on benchmarking ML with over 70 participating organizations. He is also a consultant in the Facebook Oculus VR division.
His research is centered on mobile and edge-centric computing systems with a rare taste for cloud computing aspects, mostly as it pertains to edge computing. He directs the Edge Computing Lab. He believes in solving computing problems, rather than associating himself with a particular domain or field of computing (i.e., hardware or software).
