High Dimensional and Bayesian Inference toward Quantifying Real-World Uncer...

Event Information

Share this event

Date and Time

Location

Location

Hepple Lecture Theatre

School of Geographical Sciences

University of Bristol

Bristol

BS8 1RL

United Kingdom

View Map

Event description

Description

A new collaborative event between the University of Bristol and The Institute of Statistical Mathematics in Japan.

Quantifying real-world uncertainties often come with huge challenges.

We need to use limited amount of information to infer the behaviours of a large number of random variables. Our data points may be generated from non-stationary sources due to spatial and temporal variations. However, with recent developments in high dimensional statistics, efficient Bayesian estimation and non-stationary data analysis, various theories and methodologies have been proposed to address these issues and promising results have been reported.

In this two day workshop, we have invited theoretical experts from (interpretable) high dimensional statistics, Bayesian estimation and change-point detection. Together with domain experts from bioinformatics, computational neurosciences at both the University of Bristol and The Institute of Statistical Mathematics in Japan, they will present their latest work on quantifying and exploring uncertainties of large-scale datasets in changing environments.

University of Bristol speakers:

  • Christophe Andrieu: On some properties of some Monte Carlo methods based on nonreversible processes
  • Andrew Dowsey: High-dimensional sparse inference and Bayesian modelling for finding trace-level biomarkers in large-scale mass spectrometry proteomics data
  • Peter Flach: Performance Evaluation in Machine Learning: The Good, the Bad, the Ugly and the Way Forward
  • Conor Houghton: A Kozachenko-Leonenko Estimator for Mutual Information on Metric Spaces
  • Yi Yu: Optimal change point detection and localisation

The Institute of Statistical Mathematics, Japan speakers:

  • Tomoko Matsui: Climate change mitigation management using Q-learning
  • Daichi Mochihashi: High-dimensional Motion Segmentation with semi-Markov Latent Gaussian processes
  • Daisuke Murakami: Spatial regression modelling for large dataset: A pre-compression
    approach
  • Ayaka Sakata: Model selection with piecewise continuous nonconvex penalty
  • Genta Ueno: Bayesian Estimation of the Observation-error Covariance and Its Application to Particle Filtering
  • Kenji Fukumizu

Any questions? Please contact jgi-coordinator@bristol.ac.uk

Share with friends

Date and Time

Location

Hepple Lecture Theatre

School of Geographical Sciences

University of Bristol

Bristol

BS8 1RL

United Kingdom

View Map

Save This Event

Event Saved