AutoBayes: Automated Bayesian Graph Exploration for Nuisance-Robust Inference

TL;DR

AutoBayes is an automated Bayesian framework that explores various graphical models to optimize nuisance-invariant representations and improve transfer learning in machine learning tasks.

Contribution

It introduces an automated approach to explore and optimize graphical models for nuisance-invariant learning and disentangled representations.

Findings

Significant performance improvements with ensemble learning across models.

Effective subject-transfer learning demonstrated.

Framework enables learning disentangled, nuisance-invariant features.

Abstract

Learning data representations that capture task-related features, but are invariant to nuisance variations remains a key challenge in machine learning. We introduce an automated Bayesian inference framework, called AutoBayes, that explores different graphical models linking classifier, encoder, decoder, estimator and adversarial network blocks to optimize nuisance-invariant machine learning pipelines. AutoBayes also enables learning disentangled representations, where the latent variable is split into multiple pieces to impose various relationships with the nuisance variation and task labels. We benchmark the framework on several public datasets, and provide analysis of its capability for subject-transfer learning with/without variational modeling and adversarial training. We demonstrate a significant performance improvement with ensemble learning across explored graphical models.