John Armstrong

• Information geometry. A family of probability distributions can be viewed as a submanifold of the space of all probability distributions. The Fisher information metric gives this the structure of a Riemannian manifold. One can then apply geometric notions such as projection, connections and curvature to diverse problems in probability, statistics and optimization.
• Pensions. Applying stochastic modelling techniques to pensions investments. For example, using such techniques to investigate the UK governments auto enrolment policy.
• Twistor Geometry. Twistor geometry relates the conformal geometry of 4 manifolds to the complex geometry of complex 3 manifolds. What happens to the objects of classical geometry in CP3 when they are viewed through the lens of conformal geometry?

John is a lecturer in financial mathematics, probability in statistics.

John has extensive industry experience in financial computing. He co-founded Yolus, developed their industry leading risk management system and was an Executive Director in Operations Technology at Goldman Sachs.

Before working in finance John was a lecturer at Trinity College Oxford. He completed a doctorate in differential geometry at Wadham College, Oxford under the supervision of Simon Salamon following an BA in mathematics, also at Wadham.

Information geometry, stochastic filtering, pensions, twistor geometry.

Personal website