Adaptive deep learning for embedded devices

Overview

Deep learning (DL) is the key technique in modern artificial intelligence (AI) that has provided state-of-the-art accuracy on many applications.

Today, although most of the computational loads of DL systems are still spent running neural networks in data centres, the ubiquity of smartphones, and the upcoming availability of self-contained wearable devices for augmented reality (AR), virtual reality (VR) and autonomous robot systems are placing heavy demands on DL-inference hardware with high energy and computing efficiencies along with rapid development of DL techniques.

Recently, we have witnessed a distinct evolution in the types of DL architecture, with more sophisticated network architectures proposed to improve edge AI inference. This includes dynamic network architectures that change with each new input in a data-dependent way, where inputs and internal states are not fixed.

Although such new DL architectural concepts have demonstrated a great potential to innovate current DL techniques, they are likely to affect the type of hardware and software that will be required to deliver such capabilities efficiently in the future. 

The project

This PhD project will focus on designing adaptive deep learning software and/or hardware for resource-constrained embedded systems, which precisely addresses this challenge and proposes to design a new flexible software/hardware framework to enable adaptive support for a variety of DL algorithms on embedded devices.

This project is partially supported by the Engineering and Physical Sciences Research Council (EPSRC) funded EDGE project led by the University of Essex.

Our EDGE Project is poised make an important contribution for optimising computing and energy efficient of future AI systems, which could enable disruptive capability for new types of intelligent devices and anonymous systems in the home, workspace, and extreme environment, and provide high performance and efficient AI solutions at the edge.

About Embedded and Intelligent Systems (EIS) Laboratory

Researchers in the EIS Laboratory carry out research in the areas of Embedded Systems and System-on-Chip design with focus on security, power, performance and reliability, advanced embedded systems and processor architectures targeted for cyber physical systems, automotive/industrial, robotics, image processing, networked and distributed sensor nodes/Internet of Things and real-time critical systems.

We have successfully pioneered what is now an industry leading solution with the technology for UltraSoC and Metrarc.

Funding

A full Home fee waiver (further fee details), or equivalent fee discount for overseas students. International students will need to cover the balance of their fees.

A doctoral stipend equivalent to the Research Councils UK National Minimum Doctoral Stipend (£15,609 in 2021-22, this is reviewed every year), plus £2,500 training bursary via Proficio funding, which may be used to cover the cost of advanced skills training including conference attendance and travel.

Criteria

The successful candidate would be expected to speak fluent English and meet our English Language requirements and will have a good honours BSc or BEng degree (1st, 2:1, or equivalent) in computer science, electronic engineering or a related subject.

An MSc with Merit or Distinction is desirable (but not essential for students with a first-class degree). Strong analytical and mathematical skills are required, as well as good programming skills in C/C++ and/or Python.

Knowledge of microprocessor architecture, machine learning, field-programmable gate array (FPGA), hardware description language (HDL), and/or embedded systems are desirable but not essential.

How to apply

You can apply for this postgraduate research opportunity online.

Please include your CV, covering letter, personal statement, and transcripts of UG and Masters degrees in your application.

The University has moved to requiring only one reference for PhD applications and these can be received after a conditional offer has been made so the absence of these will not hold up the recruitment process.

Instruction to applicants

When you apply online you will be prompted to fill out several boxes in the form:

• For "Course title" please put "PhD Computing and Electronic Systems".
• For "Proposed research topic or area of research" please put the title of this studentship (Adaptive deep learning for embedded devices)
• For "If you have contacted a potential supervisor..." please put Dr Xiaojun Zhai

If you have any informal queries about this opportunity please email the supervisor, Dr Xiaojun Zhai (xzhai@essex.ac.uk).

Closing date for applications is 18th June 2022.