Training neural network ensembles via trajectory sampling
Abstract
In machine learning, there is renewed interest in neural network ensembles (NNEs), whereby predictions are obtained as an aggregate from a diverse set of smaller models, rather than from a single larger model. Here, we show how to define and train a NNE using techniques from the study of rare trajectories in stochastic systems. We define an NNE in terms of the trajectory of the model parameters under a simple, and discrete in time, diffusive dynamics, and train the NNE by biasing these trajectories towards a small time-integrated loss, as controlled by appropriate counting fields which act as hyperparameters. We demonstrate the viability of this technique on a range of simple supervised learning tasks. We discuss potential advantages of our trajectory sampling approach compared with more conventional gradient based methods.
Cite
@article{arxiv.2209.11116,
title = {Training neural network ensembles via trajectory sampling},
author = {Jamie F. Mair and Dominic C. Rose and Juan P. Garrahan},
journal= {arXiv preprint arXiv:2209.11116},
year = {2023}
}
Comments
12 pages, 5 figures, 1 appendix