Associate Professor, University of Edinburgh, University of Edinburgh
1 paper at NeurIPS 2025
We make neural network training cheaper and more accurate by progressively dropping parts of the data after each epoch.
