Deep Learning on Dynamic Networks

This project is part of a 4-year Dual PhD degree programme between the National Tsing Hua University (Taiwan) and the University of Liverpool (England). As part of the NTHU-UoL Dual PhD Award students are in the unique position of being able to gain 2 PhD awards at the end of their degree from two internationally recognised world-leading Universities. As well as benefiting from a rich cultural experience, students can draw on large-scale national facilities of both countries and create a worldwide network of contacts across two continents.

The latest set of projects targeted goal #11 from the UN Sustainable Development Goals: Sustainable Cities and Communities.
This project is dedicated to deep learning on dynamic networks. Deep learning has been successful in modern artificial intelligence applications, ranging from computer vison, to machine translation, robotics, to name a few. This project will learn representations from data resided in networks (e.g., social networks, Internet of Things), where data has certain graph structures due to arbitrary network connectivity. For such structured data, a variety of graph neural network (GNN) models have been proposed and successfully applied for representation learning. While most state-of-the-art GNN models were built for static networks, this project will take network dynamics into account, where network connectivity patterns are time-varying, due to the fact that (1) network participants may join or leave the network in an autonomous manner, and (2) the network connectivity pattern may change over time due to dynamic activities and interactions among network participants.

This project aims to design novel graph neural network architectures and algorithms for supervised and semi-supervised learning, making use of spectral/structural graph theory and graph signal processing. The research outcomes will be beneficial to big data analytics for social networks, network analysis for Internet of Things, and traffic control for transportation networks. This PhD project is an exciting opportunity to delve into an important and timely research area on machine learning, graph algorithms, signal processing and data analysis. The project will help expose the PhD candidate to cutting-edge theoretical developments in these research areas as well as their applications in industrial practice. A PhD within this project will prepare the candidate to undertake academic research and industrial R&D career.

We look for ambitious and highly motivated candidates with good first degree (BSc or MSc) in Electrical Engineering, Computer Science, Mathematics or closely related subject, and with a strong interest and background in algorithms and optimisation, and programming skills. A mathematically oriented mind is desired. The applicant must have good communication skills, both verbal and in writing (English), be self-motivated and helpful team member.

For academic enquires please contact [Email Address Removed]; [Email Address Removed].
For enquires on the application process or to find out more about the Dual programme please contact School of Electrical Engineering and Electronics Postgraduate Office [Email Address Removed]


Applicants should apply via the University of Liverpool application form, for a PhD in the subject area listed above at: https://www.liverpool.ac.uk/study/postgraduate-research/how-to-apply/