The NeST seminars are a series of talks, followed by informal discussions, on topics relating the study of large dynamic networks and the themes covered in the NeST programme. Each seminar is one hour long and takes place virtually, on Microsoft Teams.

Please contact Alexander Modell (a.modell@imperial.ac.uk) if you would like to join the seminar.

Seminar schedule

• Wednesday 1st May 2024, 3pm (London time)
  Anru Zhang (Duke University): High-order Singular Value Decomposition in Tensor Analysis

  Abstract: The analysis of tensor data, i.e., arrays with multiple directions, is motivated by a wide range of scientific applications and has become an important interdisciplinary topic in data science. In this talk, we discuss the fundamental task of performing Singular Value Decomposition (SVD) on tensors, exploring both general cases and scenarios with specific structures like smoothness and longitudinality. Through the developed frameworks, we can achieve accurate denoising for 4D scanning transmission electron microscopy images; in longitudinal microbiome studies, we can extract key components in the trajectories of bacterial abundance, identify representative bacterial taxa for these key trajectories, and group subjects based on the change of bacteria abundance over time. We also showcase the development of statistically optimal methods and computationally efficient algorithms that harness valuable insights from high-dimensional tensor data, grounded in theories of computation and non-convex optimization.

  Wednesday 6th November 2024, 4pm (London time)
  Riccardo Rastelli (University College Dublin): A latent position model for multivariate time series analysis

  Multivariate time series models can suffer from overparameterization and difficult interpretability of the model parameters. In this talk, I will introduce a new statistical framework where we combine some recent network models with vector autoregressive models. The network approach is based on latent variables and, in particular, on a latent space visualization of the time series data. This formulation enables a parsimonious dependency structure and it provides new model-based visualizations of the data. In addition, the temporal aspects of the time series are naturally modeled using dynamic networks models. Hence, I will go through different variants of the proposed model, characterizing some of its properties and highlighting its performance on simulated datasets. Finally, I will propose a study on the number of recorded mumps cases in England during the last decade. 

  Wednesday 12th February 2025, 4pm (London time)
  Eric Kolaczyk (McGill University): title TBC

  Wednesday 15th May 2025, 4pm (London time)
  Konstantinos Fokianos (University of Cyprus): title TBC