Adaptive Sampling and Regularization for Stochastic Trust-region Methods
Abstract
Professor Sara Shashaani is going to talk about: 'Adaptive Sampling and Regularization for Stochastic Trust-region Methods'
Trust-region methods have proven highly effective for unconstrained nonconvex stochastic optimization problems where objective and gradient information are available only through noisy stochastic oracles. ASTRO is a class of adaptive sampling trust-region methods that dynamically determine sampling effort while constructing local quadratic models from noisy function and gradient observations. By exploiting dependence among samples and the stochastic structure of the problem, ASTRO achieves strong convergence and complexity guarantees. Its derivative-free variant, ASTRO-DF, relies solely on noisy function evaluations and also enjoys almost-sure convergence guarantees.
You may be aware that best-selling science writer Simon Singh, author of books such as “Fermat’s Last Theorem”, visited the Mathematical Institute at the start of the academic year. He and his colleague, Junaid Mubeen run a maths excellence programme (Parallel) for state school students. As well as giving some insights into what constitutes good practice in writing, outreach and education, they gave a brief outline of their Parallel programme, which currently supports 1,000 talented maths students every week from Year 7 through to Year 11.
The University of Liverpool invite applicants to their MSc and MRes programmes in Data Science & AI for Health Innovation.