Model Inversion Networks for ModelBased Optimization
Abstract
In this work, we aim to solve datadriven optimization problems, where the goal is to find an input that maximizes an unknown score function given access to a dataset of inputs with corresponding scores. When the inputs are highdimensional and valid inputs constitute a small subset of this space (e.g., valid protein sequences or valid natural images), such modelbased optimization problems become exceptionally difficult, since the optimizer must avoid outofdistribution and invalid inputs. We propose to address such problem with model inversion networks (MINs), which learn an inverse mapping from scores to inputs. MINs can scale to highdimensional input spaces and leverage offline logged data for both contextual and noncontextual optimization problems. MINs can also handle both purely offline data sources and active data collection. We evaluate MINs on tasks from the Bayesian optimization literature, highdimensional modelbased optimization problems over images and protein designs, and contextual bandit optimization from logged data.
 Publication:

arXiv eprints
 Pub Date:
 December 2019
 arXiv:
 arXiv:1912.13464
 Bibcode:
 2019arXiv191213464K
 Keywords:

 Computer Science  Machine Learning;
 Statistics  Machine Learning