Feature Density Estimation for Out-of-Distribution Detection via Normalizing Flows

no code implementations • 9 Feb 2024 • Evan D. Cook, Marc-Antoine Lavoie, Steven L. Waslander

This is a post-hoc method which can be applied to any pretrained model, and involves training a lightweight auxiliary normalizing flow model to perform the out-of-distribution detection via density thresholding.

Density Estimation Out-of-Distribution Detection +1

Class Instance Balanced Learning for Long-Tailed Classification

no code implementations • 11 Jul 2023 • Marc-Antoine Lavoie, Steven Waslander

We propose a novel class instance balanced loss (CIBL), which reweights the relative contributions of a cross-entropy and a contrastive loss as a function of the frequency of class instances in the training batch.

Classification Contrastive Learning +1