Rigorous results from machine learning | Fabian Ruehle

Speaker: Prof Fabian Ruehle, Assistant Professor at the Physics and Mathematics Department of Northeastern University College of Science

Despite their successes, machine learning techniques are often stochastic, error-prone, and blackbox. How could they then be used in fields such as theoretical physics and pure mathematics for which error-free results and deep understanding are a must?
I will discuss how we can "gamify" science problems and then tackle them with Reinforcement Learning. This technique can be applied to many hard problems from discrete mathematics and can produce provably correct results for decision problems. Moreover, studying how the game is won, i.e., how problem is solved, is typically easier than trying to interpret a neural network directly. I will illustrate this idea using examples from theoretical physics and mathematics.

This talk is part of the Liverpool Virtual Seminar Series on Data Intensive Science; more information can be found at https://indico.ph.liv.ac.uk/e/data_sc...

#data #datascience #bigdata #AI