Towards Maximizing the Representation Gap between In-Domain & Out-of-Distribution Examples

2020/10/20 by Jay Nandy, Wynne Hsu, Nandy, Jay +3 · 1 citation
Computer Science · #Speech Recognition and Synthesis #Domain Adaptation and Few-Shot Learning #Music and Audio Processing

paper · pdf · doi:10.48550/arxiv.2010.10474

Abstract

Among existing uncertainty estimation approaches, Dirichlet Prior Network (DPN) distinctly models different predictive uncertainty types. However, for in-domain examples with high data uncertainties among multiple classes, even a DPN model often produces indistinguishable representations from the out-of-distribution (OOD) examples, compromising their OOD detection performance. We address this shortcoming by proposing a novel loss function for DPN to maximize the representation gap between in-domain and OOD examples. Experimental results demonstrate that our proposed approach consistently improves OOD detection performance.

Cited by

• Uncertainty-Aware Deep Learning for Wildfire Danger Forecasting

Related

• Decomposing Representations for Deterministic Uncertainty Estimation
• Uncertainty Estimation Using a Single Deep Deterministic Neural Network
• p-DkNN: Out-of-Distribution Detection Through Statistical Testing of Deep Representations
• Failure Prediction by Confidence Estimation of Uncertainty-Aware Dirichlet Networks
• Master's Thesis: Out-of-distribution Detection with Energy-based Models