Dr Beate Ehrhardt Commercial Research Associate
Beate specialises in machine learning, hypothesis testing, Bayesian statistics and optimal experimental design.
Beate is an applied mathematician and data analysis expert whose research interests include machine learning, hypothesis testing methods, Bayesian statistics, and optimal experimental design.
Recent work includes analysis of pharmaceutical data, insurance data, and social sciences data.
Academic and commercial projects Beate has worked on include:
- Classification of X-rays using neural networks
- Design and analysis of complex pharmaceutical experiments to inform decision making across the pre-clinical pipeline
- Development of a best practice standards for experiments (e.g. confocal imaging of organs-on-chips) to reduce bias and variability
- Improvement of the statistical analysis of a neurobehavioral experiment which achieved an 80% reduction of false-positives
- Network analysis of collaborations between colleagues in a commercial setting
- Hypothesis test development for the quality of communities in networks where the data are non-identical and dependent
- Network analysis of five social networks identifying covariates that represent the structure of the connections
- Identiﬁcation of customers with a discrepancy between expected and observed loss for a car insurance company
Before joining the IMI, Beate worked as a Senior Research Statistician on pre-clinical data at the global pharmaceutical company AstraZeneca. As part of the Discovery Sciences team, she designed and analysed pharmaceutical experiments across the pre-clinical pipeline, including target validation and hit identification.
Beate holds a PhD in Mathematical Statistics from UCL (University College London) and an MSc in Mathematics from University of Bremen, Germany.