Research

My research focuses on probabilistic methods for spatio-temporal modelling.

One line of research concerns partial differential equation (PDE) modelling using score-based models. We are particularly interested in how we can flexibly condition diffusion models, to be able to jointly tackle a mix of tasks at sampling time, such as forecasting and data assimilation.

Another line of research focuses on neural processes (NPs) as a flexible probabilistic framework for modelling spatio-temporal data. Here, I worked on introducing inductive biases (e.g. translation equivariance) into NP architectures, as well as scaling transformer NPs to large datasets.

You can also find my publications on my Google Scholar profile.

Publications

1. Matthew Ashman*, Cristiana Diaconu*, Eric Langezaal*, Adrian Weller, Richard E Turner
   Gridded Transformer Neural Processes for Large Unstructured Spatio-Temporal Data
   Under review
   Paper

2. Federico Bergamin*, Cristiana Diaconu*, Aliaksandra Shysheya*, Paris Perdikaris, José Miguel Hernández-Lobato, Richard E Turner, Emile Mathieu
   On conditional diffusion models for PDE simulations
   Accepted at the 2024 Conference on Neural Information Processing Systems (NeurIPS)
   Paper

3. Matthew Ashman*, Cristiana Diaconu*, Adrian Weller, Wessel Bruinsma, Richard E Turner
   Approximately Equivariant Neural Processes
   Accepted at the 2024 Conference on Neural Information Processing Systems (NeurIPS)
   Paper

4. Matthew Ashman*, Cristiana Diaconu*, Adrian Weller, Richard E Turner
   In-Context In-Context Learning with Transformer Neural Processes
   6th Symposium on Advances in Approximate Bayesian Inference (AABI), 2024
   Paper

5. Matthew Ashman, Cristiana Diaconu, Junhyuck Kim, Lakee Sivaraya, Stratis Markou, James Requeima, Wessel P Bruinsma, Richard E Turner
   Translation Equivariant Transformer Neural Processes
   Proceedings of the 41st International Conference on Machine Learning (ICML), 2024
   Paper

6. Federico Bergamin*, Cristiana Diaconu*, Aliaksandra Shysheya*, Paris Perdikaris, José Miguel Hernández-Lobato, Richard E Turner, Emile Mathieu
   Guided autoregressive diffusion models with applications to PDE simulation
   ICLR 2024 Workshop on AI4DifferentialEquations In Science
   Paper

7. Richard E Turner, Cristiana Diaconu, Stratis Markou, Aliaksandra Shysheya, Andrew YK Foong, Bruno Mlodozeniec
   Denoising Diffusion Probabilistic Models in Six Simple Steps
   arxiv, 2024
   Paper