Cambridge Service Alliance doctoral student, Tim Pearce, explains how his work is contributing to advances in neural networks and how they are starting to help organisations make better decisions.

While our understanding and application of neural networks are progressing at speed, there is still much work to be done. One of their drawbacks is that while they are very good at giving you an answer, they aren't so good at telling you how sure they are about that answer. By using advanced mathematical tools (specifically, a Bayesian framework) we are helping them to get better at recognising and quantifying their uncertainty.

In this article, we explain the theory and show how it is being applied to a very real-world problem: helping an aircraft manufacturer manage its inventory more efficiently.

Read Inside neural networks: and how they are helping organisations make better decisions