Hey, I’m Kia! I’m a DPhil student in the Department of Statistics at the University of Oxford and part of the StatML CDT. My research interests relate broadly to the following areas and questions:
- Scalable, reliable deep learning:
- How can we reliably and efficiently quantify the uncertainty in neural network predictions?
- Generalization under distribution shifts:
- How can we train models that generalize well in the presence of distribution shifts?
- Data-efficient learning:
- How can we continue to generalize in the presence of new data (continual learning)?
- How can we learn in situations where obtaining new data is expensive (active learning, semi-supervised learning)?
Currently, I’m working on deep long-tailed classification with François Caron. Previously, I worked on analyzing the quality of variational approximations to Bayesian neural networks with Mark van der Wilk as part of my Master’s at Imperial College London. Besides my interests in machine learning, I’m also interested in:
- Long-distance cycling
- MMA (Brazilian Jiu Jitsu)
- Economics and sociology