Research
Current Research Focus
- Statistical theory of unsupervised learning
- Theory of deep learning (some recent works described in this talk)
- Non-parametric methods for learning
- Machine learning on graphs
Selected projects
- Consistency of kernel clustering
- Generalisation and asymptotics of graph neural networks
- Analysis of over-parametrised neural networks
- Machine learning on ordinal data
- Testing and clustering of large networks
Codes
- Graph neural tangent kernel
- Machine learning for life cycle assessment
- Hierarchical clustering for ordinal data
- Graph two-sample testing
- Nearest neighbour search for ordinal data (also see follow-up work on random forest for ordinal data)
- Hypergraph clustering (with application to subspace clustering)