Thu 23 Apr 2026 - The Machine Learning in Reserving Working Party will present some of our latest thinking based on our neural network (NN) framework. We will discuss the work that is in our recently (or soon to be) published blogs:
We will also cover our most recent findings from looking at time series based neural networks. Using a GRU (gated recurrent unit) seems a significant step up from a basic NN on reserving data.
We hope this will encourage reserving actuaries to have a go themselves on real data. All our research has accompanying code which is available to everyone. Our series of blogs made to accompany the NN framework were designed to allow those with no prior knowledge to learn how to use NNs.
This presentation directly follows on from our General Insurance Spring Conference 2024 presentation which we would recommend as a pre-watch.
Sarah MacDonnell, LCP and Jacky Poon