Harita is a second-year PhD student at the Warwick CDT in Mathematics & Statistics under the supervision of Theo Damoulas. Prior to this, she obtained a BSc in Mathematics from the University of Warwick and an MSc in Computational Statistics & Machine Learning from UCL where she completed a thesis on Online Meta-Learning. Her current interests lie in statistical machine learning and Bayesian methods with an emphasis on robustness and model misspecification.
Harita’s research focuses on statistical machine learning and generalised Bayesian inference methods with the goal of producing robust algorithms, suitable in cases where the model is misspecified. Her recent work has been looking at robust inference methods applicable to simulator-based models with broad applications in statistics and machine learning. In such models, evaluation of the likelihood function is not possible so inference is solely based on sampling. Simulator-based models often describe, usually via a rough approximation, some complicated physical or biological phenomena and hence they can be easily misspecified in practice. Naturally, the degree to which this approximation deviates from the true data-generating mechanism can lead to misleading inference outcomes.
Motivated by such challenges she has developed a keen interest for robust methodologies and associated algorithms suitable for modern, large-scale applications. Harita is further interested in exploring how such inferential strategies, can assist modelling in healthcare and clinical settings.