Why Should we Combine Training and Post-Training Methods for Out-of-Distribution Detection?
This work addresses the issue of unreliable OOD detection in deep learning, which is critical for safety-critical applications, but it is incremental as it builds on existing methods through combination.
The paper tackles the problem of deep neural networks making overconfident predictions for out-of-distribution (OOD) examples by reviewing recent OOD detection algorithms, categorizing them into training and post-training methods, and experimentally demonstrating that combining these approaches achieves state-of-the-art results in OOD detection.
Deep neural networks are known to achieve superior results in classification tasks. However, it has been recently shown that they are incapable to detect examples that are generated by a distribution which is different than the one they have been trained on since they are making overconfident prediction for Out-Of-Distribution (OOD) examples. OOD detection has attracted a lot of attention recently. In this paper, we review some of the most seminal recent algorithms in the OOD detection field, we divide those methods into training and post-training and we experimentally show how the combination of the former with the latter can achieve state-of-the-art results in the OOD detection task.