Overview
Yarin Gal is a machine-learning researcher (University of Oxford; AI safety and uncertainty), known for Bayesian deep learning and uncertainty quantification in neural networks.
Key Contributions / Features
- Monte Carlo dropout (Gal-Ghahramani 2016): dropout as approximate Bayesian inference, giving uncertainty from standard networks.
- Work on Concrete/variational dropout, active learning with deep models, and reliable/trustworthy AI.
Related