Daron Anderson is a Research Fellow at CONNECT at Trinity College Dublin

His current project is about modifying existing algorithms for Online Convex Optimisation to take into account when there are predictions available for the future.

In the standard problem setup we must select an action every day, and only learn tomorrow whether that action was a good one. For example we choose which stocks to buy today, and then observe how prices change to see if we made a good choice. There are many algorithms to do this effectively with no knowledge of the future. The goal is to improve these methods when there is partial knowledge available.

Daron earned his PhD in Pure Mathematics from NUI Galway in 2019. His thesis is about the various ways certain high-dimensional objects, for example the limiting paths of a weather system, can come apart when poked with a pin. The first three of six chapters have already been published in the journal Topology and its Applications

He has presented at the 2015 Annual Summer Conference on Topology and its applications; the 2018 Galway Topology Colloquium in Birmingham; and is an invited speaker at the 2019 Summer Conference on Topology and its applications in Johannesburg.