Graph Research Lab @ ANU (https://graphlabanu.github.io/website/) aims to investigate graph-related problems by marrying the best of two worlds: traditional graph algorithms and new machine learning techniques to bridge the gap between combinatorial generalization and deep learning.
Graph Research Lab offers the following projects:
(1) Graph Deep Learning - Theory and Practice
Graph deep learning aims to apply deep learning techniques to learn from complex data that is represented as graph. In recent years, due to the rising trends in network analysis and prediction, Graph Neural Networks (GNNs) as a powerful deep learning approach have been widely applied in various fields, e.g., object recognition, image classification, and semantic segmentation. However, graphs are in irregular non-Euclidean domains. This brings up the challenge of how to design deep learning techniques in order to effectively extract useful features from (possibly arbitrary) graphs. This project aims to look into recent progress of the theoretical foundations of graph deep learning to explore the connections among the representation power, generalisation ability, and optimisation performance of GNNs, as well as various graph properties.
Students are required to have a solid background in mathematics and computer science. ANU students are expected to have finished COMP4670 or have an equivalent experience.
Background literature:
- A New Perspective on “How Graph Neural Networks Go Beyond Weisfeiler-Lehman?”, A. Wijesinghe and Q. Wang, ICLR 2022.
- Future Directions in Foundations of Graph Machine Learning, Morris et al., arXiv 2024.
(2) Scalable Graph Algorithms for Data Science
Graph algorithms provide fundamental and powerful ways of exploiting the structure of graphs. In today’s real-world applications, graphs are ubiquitously used for representing complex objects and their relationships such as cities in a road network, atoms in a molecule, friendships in social networks, connections in computer networks, and links among web pages. This project aims to develop new techniques and algorithms to advance state-of-the-art research in the area of graph algorithms, graph theory, and network science.
Students are required to have a solid background in algorithms and excellent programming skills in C/C++/Java. ANU students are expected to have finished COMP3600 or have an equivalent experience.
Background literature:
- Hierarchical Cut Labelling - Scaling Up Distance Queries on Road Networks, Farhan et al., SIGMOD 2024.
- Query-by-Sketch: Scaling Shortest Path Graph Queries on Very Large Networks, Ye et al., SIGMOD 2021.