Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 2 of 2 for “"Escaping saddle points"”.
-
Noisy Iterative Learning Algorithms: Optimization, Generalization, and Privacy
… non-convex and contains local minima, saddle points, and high-dimensional, which makes the training process difficult. Second, many applications involve training data with highly sensitive information about individuals. The rich representation of deep neural networks can potentially …
-
Complexity, conditioning, and saddle avoidance in nonsmooth optimization
… guarantees, we turn to the question of escaping saddle points of nonsmooth functions. Recent work has shown that stochastically perturbed gradient methods can efficiently strict saddle points of smooth functions. We extend this body of work to nonsmooth optimization, by analyzing an …