Speaker
Michelle Kuchera
(Davidson College)
Description
Nuclear and particle physics research relies on accurate models which generate samples from conditional densities. An implicit quantile network (IQN) is a simple neural network-based machine learning model that has the ability to generate accurate samples from conditional, joint probability density functions. In this talk, we illustrate the capabilities of IQNs for simple generative tasks, as well as for the physics context of jet simulations. Specifically, we emulate folding, a stochastic process where a jet is smeared by a response function, such as interactions with a detector.
Primary author
Michelle Kuchera
(Davidson College)
