Physical Perspectives on Modern Computation

Contemporary approaches to tasks in computation (e.g. machine learning, optimisation, Bayesian statistics) draw rich and varied inspiration from models of physics. In this talk, I will outline some examples where this approach has proven fruitful, as well as more general conceptual links between physical thinking and (statistical) computation. I will then attempt to give a convincing overview of why such methods succeed, and why we should expect them to continue to offer insight into such tasks going forward.

• Speaker: Samuel Power (University of Cambridge)
• Wednesday 14 June 2017, 16:00–17:00
• Venue: MR14, Centre for Mathematical Sciences.
• Series: Cambridge Analysts’ Knowledge Exchange (C.A.K.E.); organiser: Nicolai Baldin.