Out of Distribution Reasoning by Weakly-Supervised Disentangled Logic Variational Autoencoder

TL;DR

This paper introduces a framework for out-of-distribution reasoning using a partially disentangled VAE trained with weak supervision, enabling better detection and reasoning about complex, unknown generative factors in datasets like Carla.

Contribution

It proposes a novel three-step framework that combines data partitioning, logic tensor network training, and runtime reasoning for complex datasets with unknown factors.

Findings

Outperformed state-of-the-art methods in disentanglement.

Achieved superior OOD reasoning on the Carla dataset.

Demonstrated effectiveness of partial disentanglement in complex scenarios.

Abstract

Out-of-distribution (OOD) detection, i.e., finding test samples derived from a different distribution than the training set, as well as reasoning about such samples (OOD reasoning), are necessary to ensure the safety of results generated by machine learning models. Recently there have been promising results for OOD detection in the latent space of variational autoencoders (VAEs). However, without disentanglement, VAEs cannot perform OOD reasoning. Disentanglement ensures a one- to-many mapping between generative factors of OOD (e.g., rain in image data) and the latent variables to which they are encoded. Although previous literature has focused on weakly-supervised disentanglement on simple datasets with known and independent generative factors. In practice, achieving full disentanglement through weak supervision is impossible for complex datasets, such as Carla, with unknown and abstract generative factors. As a result, we propose an OOD reasoning framework that learns a partially disentangled VAE to reason about complex datasets. Our framework consists of three steps: partitioning data based on observed generative factors, training a VAE as a logic tensor network that satisfies disentanglement rules, and run-time OOD reasoning. We evaluate our approach on the Carla dataset and compare the results against three state-of-the-art methods. We found that our framework outperformed these methods in terms of disentanglement and end-to-end OOD reasoning.