Ricky T. Q. Chen, Yulia Rubanova, Jesse Bettencourt, & David Duvenaud (2018)
arXiv.
DOI: https://doi.org/10.48550/arxiv.1806.07366
Abstract. Introduces Neural ODEs, which define a continuous-time analogue of residual networks and enable continuous normalising flows whose log-density is governed by an instantaneous change-of-variables formula.
Tags: generative normalising-flows neural-ode