Course 11: Bayesian Methods and Machine Learning
The TALENT/INT course on "Learning from Data: Bayesian Methods and Machine Learning" was held at the University of York in York, UK from June 10 to 28, 2019.
See the github course web page and repository for materials, including many Jupyter notebooks.
Course organizers: Christian Forssén (Chalmers University, Sweden), Dick Furnstahl (Ohio State University, USA), Daniel Phillips (Ohio University, USA), Ian Vernon (Durham University, UK)
In recent years there has been an explosion of interest in the use of Bayesian methods in nuclear physics. These methods are being used to quantify the uncertainties in theoretical work on topics ranging from the NN force to high-energy heavy-ion collisions, to develop more reliable extrapolants of nuclear-energy-density functionals towards the dripline, to predict the impact that future NICER observations may have on the equation of state of neutron matter, and to determine whether or not nucleon resonances are present in experimental data. Meanwhile machine learning is gaining increased currency as a method for identifying interesting signals in both experiments and simulations.
While most nuclear-physics Ph.D. students are taught some standard (frequentist) statistics as part of their course work, very few encounter Bayesian methods until they are engaged in research. But Bayesian methods provide a coherent and compelling framework to think about inference, and so can be applied to many important questions in nuclear physics. The overall learning goal of this school is to take students who have had no previous exposure to Bayes’ theorem and show them how it can be applied to problems of parameter estimation, model selection, and machine learning.