Introducing 'Inside' Out of Distribution
This work addresses the need for more reliable OOD detection methods in ML by highlighting a previously neglected aspect, though it appears incremental as it builds on existing OOD studies.
The paper tackles the problem of out-of-distribution (OOD) detection in machine learning by introducing a distinction between inside (interpolatory) and outside (extrapolatory) OOD cases, showing that outside OOD generally causes greater performance degradation, with analysis based on normalized RMSE and F1 scores on synthetic datasets.
Detecting and understanding out-of-distribution (OOD) samples is crucial in machine learning (ML) to ensure reliable model performance. Current OOD studies primarily focus on extrapolatory (outside) OOD, neglecting potential cases of interpolatory (inside) OOD. In this study, we introduce a novel perspective on OOD by suggesting it can be divided into inside and outside cases. We examine the inside-outside OOD profiles of datasets and their impact on ML model performance, using normalized Root Mean Squared Error (RMSE) and F1 score as the performance metrics on syntetically-generated datasets with both inside and outside OOD. Our analysis demonstrates that different inside-outside OOD profiles lead to unique effects on ML model performance, with outside OOD generally causing greater performance degradation, on average. These findings highlight the importance of distinguishing between inside and outside OOD for developing effective counter-OOD methods.