Latent Discriminant deterministic Uncertainty
About
Predictive uncertainty estimation is essential for deploying Deep Neural Networks in real-world autonomous systems. However, most successful approaches are computationally intensive. In this work, we attempt to address these challenges in the context of autonomous driving perception tasks. Recently proposed Deterministic Uncertainty Methods (DUM) can only partially meet such requirements as their scalability to complex computer vision tasks is not obvious. In this work we advance a scalable and effective DUM for high-resolution semantic segmentation, that relaxes the Lipschitz constraint typically hindering practicality of such architectures. We learn a discriminant latent space by leveraging a distinction maximization layer over an arbitrarily-sized set of trainable prototypes. Our approach achieves competitive results over Deep Ensembles, the state-of-the-art for uncertainty prediction, on image classification, segmentation and monocular depth estimation tasks. Our code is available at https://github.com/ENSTA-U2IS/LDU
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Semantic segmentation | CholecSeg8K (test) | -- | 13 | |
| 3D Object Detection | aiMotive (test) | Mean AP (all-point)64.69 | 12 | |
| Tissue Segmentation | CholecSeg8K corrupted (test) | Precision1.44 | 11 | |
| Tissue Segmentation | Endoscopic Submucosal Dissection (ESD) (test) | ECE14.12 | 11 | |
| Tissue Segmentation | Endoscopic Submucosal Dissection (ESD) OOD corrupted (test) | PR Score1.33 | 11 | |
| 3D Object Detection | aiMotive Over-Exposure | AP (all-point)65.07 | 6 | |
| 3D Object Detection | aiMotive LiDAR-Fog | AP (All-Point)65.47 | 6 | |
| 3D Object Detection | aiMotive distant region >75m (test) | Highway (all-point AP)43.07 | 6 | |
| 3D Object Detection | aiMotive Motion Blur | AP (All-Point)62.06 | 6 | |
| 3D Object Detection | aiMotive Under-Exposure | AP (All-Point)62.87 | 6 |