Optimal Transport-Induced Samples against Out-of-Distribution Overconfidence

TL;DR

This paper introduces a novel method using optimal transport to generate semantically ambiguous OOD samples, which helps calibrate neural networks and reduce overconfidence on out-of-distribution inputs.

Contribution

The paper proposes a new OT-based framework to identify and sample near semantic boundaries, improving model calibration against OOD overconfidence.

Findings

Significantly reduces OOD overconfidence in neural networks.

Outperforms existing methods in calibration and robustness.

Provides a geometrically grounded approach to OOD detection.

Abstract

Deep neural networks (DNNs) often produce overconfident predictions on out-of-distribution (OOD) inputs, undermining their reliability in open-world environments. Singularities in semi-discrete optimal transport (OT) mark regions of semantic ambiguity, where classifiers are particularly prone to unwarranted high-confidence predictions. Motivated by this observation, we propose a principled framework to mitigate OOD overconfidence by leveraging the geometry of OT-induced singular boundaries. Specifically, we formulate an OT problem between a continuous base distribution and the latent embeddings of training data, and identify the resulting singular boundaries. By sampling near these boundaries, we construct a class of OOD inputs, termed optimal transport-induced OOD samples (OTIS), which are geometrically grounded and inherently semantically ambiguous. During training, a confidence suppression loss is applied to OTIS to guide the model toward more calibrated predictions in structurally uncertain regions. Extensive experiments show that our method significantly alleviates OOD overconfidence and outperforms state-of-the-art methods.