Clearing 2021
News

Computer scientist wins prestigious EPSRC New Investigator Award

  • Date

    Tue 6 Apr 21

Dr Xiaojun Zhai

Computer scientist Dr Xiaojun Zhai has won the prestigious Engineering and Physical Sciences Research Council’s (EPSRC) New Investigator Award.

Deep learning is the key technique in modern artificial intelligence (AI) that has provided state-of-the-art accuracy on many applications. However, due to the nature of its computational intensity it is not suitable to be deployed on resource-constrained embedded devices such as mobile phones, drones and mobile robots.

Dr Zhai, from Essex’s Robotics and Embedded Systems research group, will investigate a new adaptive hardware architecture to enable a variety of deep learning algorithms to be used on embedded devices in real-world mission-critical applications, ranging from security surveillance to self-driving robots and cars.

This research will provide high performance and efficient AI solutions for new types of intelligent devices and anonymous systems in the home, workspace and in extreme environments.

This project will work closely with Essex’s industry and project partners, including Arm, a world leading company for automotive and Internet of Things business and semiconductor company Xilinx to create a number of scenarios to showcase the effectiveness of the adaptive deep learning framework versus traditional approaches in practical settings.

As robotics and AI are key emerging technologies Dr Zhai will also work collaboratively with the National Centre for Nuclear Robotics on the development of advanced robot vision systems for nuclear decommissioning, as he believes his new research could significantly improve the energy, efficiency and reliability of future robot vision systems.

Dr Zhai said: “I am very excited to receive this New Investigator Award and it will represent a fundamental step in my long-term career plan to establish a specialist research group on hardware-efficient architectures for machine-learning algorithms in resource-constrained devices at Essex.”