Large Class Separation is not what you need for Relational Reasoning-based OOD Detection

TL;DR

This paper investigates the use of relational reasoning pre-training for out-of-distribution detection, highlighting the importance of class separation in embedding space and proposing a new loss function to improve OOD detection without additional training.

Contribution

It introduces an analysis of relational reasoning pre-training for OOD detection and proposes a new loss function to better control class separation in embeddings.

Findings

Inter-class feature distance correlates with OOD detection accuracy.

A new loss function improves class separation and OOD detection performance.

Relational reasoning pre-training enhances OOD detection without fine-tuning.

Abstract

Standard recognition approaches are unable to deal with novel categories at test time. Their overconfidence on the known classes makes the predictions unreliable for safety-critical applications such as healthcare or autonomous driving. Out-Of-Distribution (OOD) detection methods provide a solution by identifying semantic novelty. Most of these methods leverage a learning stage on the known data, which means training (or fine-tuning) a model to capture the concept of normality. This process is clearly sensitive to the amount of available samples and might be computationally expensive for on-board systems. A viable alternative is that of evaluating similarities in the embedding space produced by large pre-trained models without any further learning effort. We focus exactly on such a fine-tuning-free OOD detection setting. This works presents an in-depth analysis of the recently introduced relational reasoning pre-training and investigates the properties of the learned embedding, highlighting the existence of a correlation between the inter-class feature distance and the OOD detection accuracy. As the class separation depends on the chosen pre-training objective, we propose an alternative loss function to control the inter-class margin, and we show its advantage with thorough experiments.