New Techniques in Optimization and Their Applications to Deep Learning and Related Inverse Problems

Stanley Osher
University of California, Los Angeles (UCLA)
Mathematics

We will draw from some of our new results in optimization, often related to partial differential equations, to improve performance of algorithms ranging from data dependent activation in deep learning, training quantized neural networks, optimizing neural networks solving the phase lift problem and diagnosing forward operator error.
(joint with many people)

Presentation (PDF File)

Back to New Deep Learning Techniques