Catalyst: Out-of-Distribution Detection via Elastic Scaling

TL;DR

Catalyst introduces a novel post-hoc method that leverages raw pre-pooling feature statistics to improve out-of-distribution detection, significantly boosting existing methods' performance.

Contribution

It exploits raw feature map statistics for elastic scaling, enhancing the effectiveness of various OOD detection methods without retraining.

Findings

Reduces false positive rates by up to 32.87% on CIFAR-10

Improves OOD detection performance across multiple benchmarks

Seamlessly integrates with existing logit-based and distance-based detectors

Abstract

Out-of-distribution (OOD) detection is critical for the safe deployment of deep neural networks. State-of-the-art post-hoc methods typically derive OOD scores from the output logits or penultimate feature vector obtained via global average pooling (GAP). We contend that this exclusive reliance on the logit or feature vector discards a rich, complementary signal: the raw channel-wise statistics of the pre-pooling feature map lost in GAP. In this paper, we introduce Catalyst, a post-hoc framework that exploits these under-explored signals. Catalyst computes an input-dependent scaling factor ($\gamma$) on-the-fly from these raw statistics (e.g., mean, standard deviation, and maximum activation). This $\gamma$ is then fused with the existing baseline score, multiplicatively modulating it -- an $\textit{elastic scaling}$ -- to push the ID and OOD distributions further apart. We demonstrate Catalyst is a generalizable framework: it seamlessly integrates with logit-based methods (e.g., Energy, ReAct, SCALE) and also provides a significant boost to distance-based detectors like KNN. As a result, Catalyst achieves substantial and consistent performance gains, reducing the average False Positive Rate by 32.87 on CIFAR-10 (ResNet-18), 27.94% on CIFAR-100 (ResNet-18), and 22.25% on ImageNet (ResNet-50). Our results highlight the untapped potential of pre-pooling statistics and demonstrate that Catalyst is complementary to existing OOD detection approaches. Our code is available here: https://github.com/bingabid/Catalyst