Jingwei Liang, University of Cambridge
601 Pao Yue-Kong Library
Stochastic optimisation methods have become increasingly popular over the past few years, especially owing to the burst of machine learning. However, despite the popularity, there are still many unknowns about these methods. In this talk, we provide a local perspective to understand stochastic optimisation methods, for example the difference between different stochastic methods, connections to the deterministic counterparts, and more importantly possible ways to improve these methods.