Fun tricks in non-convex optimisation
2014-10-04 — 2022-07-14
Wherein the role of initialization and symmetries in steering final optima is examined, and phase retrieval is treated while a double-toboggan illustration is employed to show dependence on basins of attraction.
functional analysis
optimization
statmech
- Zeyuan ALLEN-ZHU: Recent Advances in Stochastic Convex and Non-Convex Optimization. Clear, has good pointers.
1 With symmetries
Zhang, Qu, and Wright (2022)
2 In phase retrieval
See phase retrieval.
3 References
Azizian, Iutzeler, Malick, et al. 2025. “The Global Convergence Time of Stochastic Gradient Descent in Non-Convex Landscapes: Sharp Estimates via Large Deviations.”
Choromanska, Henaff, Mathieu, et al. 2015. “The Loss Surfaces of Multilayer Networks.” In Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics.
Jain, and Kar. 2017. Non-Convex Optimization for Machine Learning.
Soltanolkotabi, Javanmard, and Lee. 2019. “Theoretical Insights Into the Optimization Landscape of Over-Parameterized Shallow Neural Networks.” IEEE Transactions on Information Theory.
Wright, and Ma. 2022. High-dimensional data analysis with low-dimensional models: Principles, computation, and applications.
Zhang, Qu, and Wright. 2022. “From Symmetry to Geometry: Tractable Nonconvex Problems.”