Track 4: Deploying AI in the Enterprise

Evolution of AI Chip Architectures to the Edge


Salon G December 13, 2017 2:20 pm - 3:00 pm

Bookmark and Share

Shriram Ramanathan, Ph.D.
Oskar Mencer
Dinakar Munagala
Shinji Hayakawa
Jeff Burns

Moderator: Dr. Shriram Ramanathan, Senior Analyst, Lux Research

Panelists:

– Shinji Hayakawa, Technical Communicator, Semiconductor Energy Laboratory, Co., Ltd

– Oskar Mencer, Founder, Maxeler Technologies

– Dinakar Munagala, CEO and Founder, ThinCI

– Jeff Burns, Director, Systems Architecture and Design, IBM Research

As companies adopt AI and advanced analytics to leverage big data, many are rethinking the widely prevalent centralized cloud analytics paradigm. Large companies such as Nvidia, Intel, and Google as well as startup companies are developing novel and powerful chip architectures that will eventually move AI from centralized, cloud-based systems to devices at the edge of the network.

In this panel, we will discuss the reasons why new chip architectures are essential in order to enable edge computing and fully unleash the power of AI, and in particular, to bring forth new applications and use cases. Importantly, we will highlight some recent AI chip architectures, their capabilities as well as the key challenges in designing such architectures.