Eshed Ohn Bar—Scaling Systems with Everyone, Everywhere, All the Time: The Co Ops Project скачать в хорошем качестве

Eshed Ohn Bar—Scaling Systems with Everyone, Everywhere, All the Time: The Co Ops Project

Wednesday, February 28, 2024 2024 MOC Alliance Workshop Session 1—AI and the AI Alliance Speaker: Eshed Ohn-Bar, Assistant Professor, Boston University Talk Title: Scaling Systems with Everyone, Everywhere, All the Time: The Co-Ops Project Talk Abstract: Independent and closed-source development of societally beneficial AI-based edge systems, such as autonomous vehicles, has been frustratingly slow, costly, and inefficient. In contrast, collaborative development and training of large AI models present an opportunity to rapidly advance scalable, privacy-preserving, and more transparent edge systems today. In this talk, I will introduce the Co-Ops project, a new framework for robust and efficient distributed learning from highly diverse data. The framework significantly enhances computational efficiency and system performance (by over 17%), moving beyond the limitations of traditional Federated Learning and secure Multi-Party Computation. I will then present a set of Red Hat-integrated, open-source tools that facilitate the collection, protection, and processing of unconstrained data from heterogenous participants. Towards a future of collaborative AI-based systems, the introduced tools fill in a current gap in seamlessly integrating data from diverse agents at an unprecedented scale. Bio: Eshed Ohn-Bar is an Assistant Professor in the ECE department at Boston University. Prior to joining BU he was a Humboldt Fellow at the Max Planck Institute. Eshed’s research lies at the intersection of machine intelligence, systems, and accessibility. His work won the semi-finalist for the 2022 Department of Transportation’s Inclusive Design Grand Challenge and the 2017 best PhD dissertation award from the IEEE Intelligent Transportation Systems Society. Eshed received the BS degree in Mathematics from UC Los Angeles in 2010, MEd from UC Los Angeles in 2011, and the PhD degree in Electrical Engineering from UC San Diego in 2017. Slides: https://docs.google.com/presentation/...