Panel IV: The Role of Optics in Future Machine Learning Architectures

Wednesday, 09 March 15:30 – 17:00
Theater I

Session Description:

In the mission to expand Machine Learning (ML) capabilities, system scaling/optimization and relieving key I/O bottlenecks become critical. The panel will discuss various architectures and approaches to scaling ML, and the potential role of optical technologies: from tighter integration, higher speed interfaces, to lower latency channels/compute. We will also discuss the overlap or bifurcation of these optical technologies for ML versus other high volume datacenter applications, and implications to cost and the industry as a whole.

Moderator: 

Ryohei Urata, Technical Lead/Manager, Google, USA

Panelists: 

Paolo Costa, Principal Researcher, Microsoft, USA
Optical Networking for Machine Learning: The Light at the End of the Tunnel?

Larry Dennison, Director of Network Research, Nvidia, USA
Title to be Determined

Nick Harris, CEO, Lightmatter, USA
Beyond Toy Platforms: Taking Photonic Computing to Production Workloads and Environments

Robert Stone, Technical Sourcing Manager, Meta (formerly Facebook), USA
Optical Interconnects for AI / ML Fabrics

Biographies: 

Paolo Costa, Principal Researcher, Microsoft, USA
Paolo is a Principal Researcher in the Cloud Infrastructure Group in Microsoft Research and an Honorary Lecturer with the Department of Computing of Imperial College London. His current research brings together hardware, optics, networking, and application-level expertise to take a cross-stack view towards developing optical technologies for next-generation data-center networks.

Larry Dennison, Director of Network Research, Nvidia, USA
Dr. Dennison joined NVIDIA in September of 2013 and leads the Network Research Group.  His current research interests include large networks of GPUs, switch micro-architectures, network-on-chip and photonic interconnects.  At, NVIDIA, he was the principal investigator for the DesignForward project which was responsible for several GPU shared-memory concepts such as NVSHMEM and NCCL.  His team proposed development of a GPU shared memory fabric and developed the first NVSwitch architecture. Prior to NVIDIA, he worked on software systems such as high-performance distributed applications, database scaling for the cloud and software-defined networking. He also architected and led the development of the ASIC chipset for the Avici Terabit Router which utilized a 3-D toroidal network. At BBN, Dr. Dennison was the principal investigator for MicroPathfinder, a wearable computer that connected to other wearables over a very low power RF network. Dr. Dennison holds Ph.D., M.S., and B.S. degrees from the Massachusetts Institute of Technology.

Nick Harris, CEO, Lightmatter, USA
With over 30 patents and 70 publications in journals including Nature, Nature Photonics, Nature Physics, and Physical Review X, Nicholas’ seminal work on quantum and classical information processing with integrated photonics has helped launch the international research field of programmable photonics. Nicholas has been recognized by the MIT Technology Review through the prestigious Innovators Under 35 award (2021). The award had over 500 nominees and a panel of 30 judges including artificial intelligence luminary Andrew Ng; previous winners include Larry Page and Sergey Brin (Google founders), Mark Zuckerberg (Facebook founder), and Jony Ive (Apple). He received his doctorate in Electrical Engineering and Computer Science at the Massachusetts Institute of Technology where he was a National Science Foundation Graduate Research Fellow and later served as a Postdoctoral Fellow through the Intelligence Community Postdoctoral Fellowship.

Robert Stone, Technical Sourcing Manager, Meta (formerly Facebook), USA
Rob Stone is a member of the infrastructure team at Meta, focusing on next generation optics and networking. Prior to Facebook, Rob was a distinguished engineer at Broadcom within the switch architecture team and is active in a of industry communities including IEEE as well as MSAs and industry organizations. Rob holds a D.Phil. in Physics from The University of Oxford.

Ryohei Urata, Technical Lead/Manager, Google, USA
Dr. Ryohei Urata is currently a Principal Engineer/Director in the Platforms Optics Group, where he has defined/developed Google's datacenter optical technologies and corresponding roadmap for the past decade. Prior to joining Google, he was a research specialist at NTT Photonics Laboratories, Japan. He has over 150 patents, publications, and presentations in the areas of optical interconnect, switching, and networking. He received the B.S. degree in engineering physics from the University of California at Berkeley (Highest Honors), and the M.S. and Ph.D. degrees in electrical engineering from Stanford University (Stanford Graduate Fellow). He was elected Optica/OSA Fellow in 2022.