Enhancing Out-of-Distribution Detection with Extended Logit Normalization
This addresses the need for safe deployment of machine learning models by enhancing out-of-distribution detection, though it appears incremental as it builds on Logit Normalization.
The paper tackled the problem of out-of-distribution detection by identifying a critical issue in Logit Normalization and proposing Extended Logit Normalization, which significantly improves a wide range of post-hoc detection methods and outperforms state-of-the-art training-time methods in benchmarks.
Out-of-distribution (OOD) detection is essential for the safe deployment of machine learning models. Recent advances have explored improved classification losses and representation learning strategies to enhance OOD detection. However, these methods are often tailored to specific post-hoc detection techniques, limiting their generalizability. In this work, we identify a critical issue in Logit Normalization (LogitNorm), which inhibits its effectiveness in improving certain post-hoc OOD detection methods. To address this, we propose Extended Logit Normalization ($\textbf{ELogitNorm}$), a novel hyperparameter-free formulation that significantly benefits a wide range of post-hoc detection methods. By incorporating feature distance-awareness to LogitNorm, $\textbf{ELogitNorm}$ shows more robust OOD separability and in-distribution (ID) confidence calibration than its predecessor. Extensive experiments across standard benchmarks demonstrate that our approach outperforms state-of-the-art training-time methods in OOD detection while maintaining strong ID classification accuracy.