Deep Quantization for Energy Efficient Inference at the Edge

Wednesday, May 23, 10:40 AM - 11:10 AM
Summit Track: 
Technical Insights I
Mission City Ballroom B2-B5

Intelligence at the edge is different from intelligence in the cloud in terms of requirements for energy, cost, accuracy and latency. Due to limits on battery power and cooling systems in edge devices, energy consumption is strictly limited. In addition, low cost and small size requirements make it hard to use packages with large numbers of pins, thus limiting the bandwidth to DRAM chips commonly used for storing neural network algorithm information. Despite these limitations, most applications require real-time operation. To tackle this issue, we have developed networks that heavily rely on deep quantization.

In this talk, we show how to use the deep quantization in real applications without degrading accuracy. Specifically, we explain the use of different quantizations for each layer of a deep neural network and how to use deep layered neural networks along with deep quantization. We also explain the use of this deep quantization approach with recent lightweight networks.


Hoon Choi

Senior Director of Design Engineering, Lattice

Hoon Choi is a Design Engineer Senior Director at Lattice Semiconductor, where he has been involved in the specification and design of multiple generations of HDMI for more than 13 years. In addition, Choi has led the design of HDCP 2.2 for the HDMI and compliance test specification. Also, he led the specification and implementation of UCP (China-CP) and the authentication implementation activity for USB Type-C. Prior to joining Lattice, Choi got his Ph.D from KAIST in Korea then began his career in the technology industry by working for Samsung Electronics, NeoPace Telecom, and Silicon Image, which was acquired by Lattice.

See you at the Summit! May 20-23 in Santa Clara, California!
Register today and reserve your hotel room!