My research explores the overlap between practical machine-learning methods and their underlying theory, a space where recent progress in training deep networks still depends largely on heuristics and has little theoretical backing. I aim to narrow the gap between theory and practice by:
1Explain
developing practical theory that explains why specific phenomena arise and why certain heuristics succeed during deep-learning training
2Improve
advancing existing algorithms from first principles
3Create
creating new scalable and efficient methods with both theoretical analysis and practical performance
