TwoTimescale Stochastic Approximation for Bilevel Optimisation Problems in ContinuousTime Models
Abstract
We analyse the asymptotic properties of a continuoustime, twotimescale stochastic approximation algorithm designed for stochastic bilevel optimisation problems in continuoustime models. We obtain the weak convergence rate of this algorithm in the form of a central limit theorem. We also demonstrate how this algorithm can be applied to several continuoustime bilevel optimisation problems.
 Publication:

arXiv eprints
 Pub Date:
 June 2022
 arXiv:
 arXiv:2206.06995
 Bibcode:
 2022arXiv220606995S
 Keywords:

 Mathematics  Optimization and Control;
 Statistics  Machine Learning
 EPrint:
 Accepted at ICML 2022 Workshop on Continuous Time Methods in Machine Learning