Gaussian Latent Representations for Uncertainty Estimation using Mahalanobis Distance in Deep Classifiers - Université de Lorraine
Conference Papers Year : 2023

Gaussian Latent Representations for Uncertainty Estimation using Mahalanobis Distance in Deep Classifiers

Abstract

Recent works show that the data distribution in a network’s latent space is useful for estimating classification uncertainty and detecting Out-Of-Distribution (OOD) samples. To obtain a well-regularized latent space that is conducive for uncertainty estimation, existing methods bring in significant changes to model architectures and training procedures. In this paper, we present a lightweight and high-performance regularization method for Mahalanobis distance (MD)-based uncertainty prediction, and that requires minimal changes to the network’s architecture. To derive Gaussian latent representation favourable for MD calculation, we introduce a self-supervised representation learning method that separates in-class representations into multiple Gaussians. Classes with non-Gaussian representations are automatically identified and dynamically clustered into multiple new classes that are approximately Gaussian. Evaluation on standard OOD benchmarks shows that our method achieves state-of-the-art results on OOD detection and is very competitive on predictive probability calibration. Finally, we show the applicability of our method to a real-life computer vision use case on microorganism classification.

Dates and versions

hal-04423169 , version 1 (29-01-2024)

Identifiers

Cite

Aishwarya Venkataramanan, Assia Benbihi, Martin Laviale, Cédric Pradalier. Gaussian Latent Representations for Uncertainty Estimation using Mahalanobis Distance in Deep Classifiers. 2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), Oct 2023, Paris, France. pp.4490-4499, ⟨10.1109/ICCVW60793.2023.00483⟩. ⟨hal-04423169⟩
38 View
0 Download

Altmetric

Share

More