A Survey on Efficient and Privacy-Preserving Deep Learning for Edge Devices

PhD Qualifying Examination


Title: "A Survey on Efficient and Privacy-Preserving Deep Learning for 
Edge Devices"

by

Mr. Han TIAN


Abstract:

The emerge of deep learning applications for our daily lives and the 
popularity of edge devices, especially IoT devices and smartphones, has 
raised the requirement of a combination of these two. For model training, 
the data generated by individuals’ edge devices can be utilized to 
collaboratively train a deep learning model. For model inference, 
deploying deep learning model on the edge can reduce response latency and 
make real-time decision making possible. However, how to perform 
computation-intense deep learning tasks efficiently on edge devices with 
limited resources, while preserving individual data privacy is 
challenging.

To resolve these challenges, during the past few years, great efforts have 
been made in this area. In this survey, we present an up-to-date and 
thorough introduction of the advances in research and industry areas 
filling the gap between deep learning and edge devices. We will discuss 
the characteristics of edge devices and deep learning models, the 
efficiency and privacy challenges raised when pushing deep learning to 
edge devices, and the various paradigms and algorithms designed for 
facilitating efficient and privacy-preserving model training and inference 
on the edge.


Date:			Monday, 26 August 2019

Time:                  	10:00am - 12:00noon

Venue:                  Room 5501
                         Lifts 25/26

Committee Members:	Prof. Qiang Yang (Supervisor)
 			Dr. Kai Chen (Supervisor)
 			Dr. Qifeng Chen (Chairperson)
 			Dr. Ming Liu (ECE)


**** ALL are Welcome ****