Equivariant Flows: Exact Likelihood Generative Learning for Symmetric Densities

Authors: Jonas Köhler, Leon Klein, Frank Noe

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Using benchmark systems motivated from molecular physics, we demonstrate that those symmetry preserving flows can provide better generalization capabilities and sampling efficiency.
Researcher Affiliation Academia 1Freie Universit at Berlin, Department of Mathematics and Computer Science. 2Freie Universit at Berlin, Department of Physics. 3Rice University, Department of Chemistry.
Pseudocode No The paper describes methods and equations, but does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statements about making its source code publicly available or include links to a code repository.
Open Datasets No The training data is generated by taking 10 / 100 / 1, 000 / 10, 000 samples from a long MCMC trajectory (throwing away 1, 000 burn-in samples to enforce equilibration). The paper defines the potential functions for the systems but does not provide concrete access information (link, DOI, etc.) for a publicly available dataset used for training.
Dataset Splits No The paper mentions generating training data and evaluating on an 'independent 10,000 trajectory' for testing, but does not specify a separate validation split for hyperparameter tuning.
Hardware Specification No The paper does not provide specific details about the hardware used, such as GPU or CPU models.
Software Dependencies No The paper mentions software like 'Adam' and 'dopri5' and 'Runge-Kutta' as solvers, but does not provide specific version numbers for any software dependencies.
Experiment Setup Yes We compare the OTD implementation presented in (Chen et al., 2018; Grathwohl et al., 2018) using the dopri5-option (atol = 10 10, rtol = 10 5) to the DTO implementation given by Gholami et al. using a fixed grid of 20 steps and 4th-order Runge-Kutta as solver. ... We train both flows using Adam with weight decay (Kingma & Ba, 2014; Loshchilov & Hutter) until convergence.