Machine learning assisted computation of the marginal likelihood — Bayesian model comparison provides a principled statistical framework for selecting an appropriate model to describe observational data, naturally trading off model complexity and goodness of fit. However, it requires computation of the Bayesian model evidence, also called the marginal likelihood, which is computationally challenging. We present the learnt harmonic mean…

Hybrid rotationally equivariant spherical CNNs — Notions of spherical convolution offer a promising route to unlocking the potential of deep learning for the variety of problems in which spherical data are prevalent. However, the introduction of non-linearity is a challenge. In this post we explore how ideas originating in quantum physics may be applied to overcome…

Spherical CNNs By encoding an understanding of the translational symmetry of the physical world, convolutional neural networks (CNNs) have revolutionised computer vision. …

How geometric AI techniques can be used to alleviate cybersickness — Today’s 360° virtual reality (VR) experiences have great potential, allowing you to be transported anywhere in the world. They are photo-realistic and relatively easy and cheap to acquire — requiring only an off-the-shelf 360° camera — but they lack immersion and interactivity. Critically, this lack of interactivity can also induce…

How to ace your PhD application interview and land that dream PhD position. — It’s that time of year when prospective PhD students are interviewing for PhD positions to start next academic year. Over the past week I’ve interviewed around 40 prospective students. Every year the calibre of applicants just seems to get better and better. I’m continually amazed by all of the cool…

Harnessing symmetry in machine learning — In many ways physics and machine learning share a common goal: to formulate models of observed phenomena. In achieving this goal physicists have long understood the importance of symmetry. …