Research map / Label-efficient and robust learning / OOD and anomaly detection
Adaptive Confidence Regularization for Multimodal Failure Detection
Moru Liu, Hao Dong, Olga Fink, Mario Trapp
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Adaptive Confidence Loss and Multimodal Feature Swapping (ACR) | MSP baseline · HMDB51 (video+flow) | AURC: 29.56 -> 19.97 | Improved |
| add Adaptive Confidence Loss only (ablation) | MSP baseline · HMDB51 (video+flow) | AURC: 29.56 -> 24.48 | Improved |
| add Multimodal Feature Swapping only (ablation) | MSP baseline · HMDB51 (video+flow) | AURC: 29.56 -> 25.11 | Improved |
| combine ACL and MFS (ablation) | ACL only · HMDB51 (video+flow) | AURC: 24.48 -> 19.97 | Improved |
| MFS instead of random noise outlier (ablation) | ACR · HMDB51 (video+flow) | AURC: 24.86 -> 19.97 | Improved |
| MFS instead of feature mixing (ablation) | ACR · HMDB51 (video+flow) | AURC: 21.79 -> 19.97 | Improved |
| ACR with alternative I3D/TSN backbones | MSP · HMDB51 | AURC: 31.77 -> 27.73 | Improved |
| ACR with audio modality combinations | MSP · HAC (video+flow+audio) | FPR95: 42.86 -> 34.78 | Improved |
| ACR applied to segmentation | MSP · SemanticKITTI | mIoU: 59.25 -> 63.56 | Improved |
| ACR vs OpenMix on EPIC-Kitchens FPR95 and accuracy | OpenMix · EPIC-Kitchens (video+flow) | FPR95: 73.68 -> 71.58 | Mixed |
| OOD detection with ACR vs A2D+NP-Mix (AN) | MSP scoring · HMDB51 ID, average over 4 OOD sets | FPR95: 27.74 -> 24.15 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- OpenMix — OpenMix: Reviving Known Knowledge for Discovering Novel Visual Categories in an Open World (CVPR 2021) · better on 10 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MultiOOD: Scaling Out-of-Distribution Detection for Multiple Modalities (NeurIPS 2024 · similarity 0.85)
- DPU: Dynamic Prototype Updating for Multimodal Out-of-Distribution Detection (CVPR 2025 · similarity 0.85)
- Balancing Multimodal Training Through Game-Theoretic Regularization (NeurIPS 2025 · similarity 0.84)
- Are We Making Progress in Multimodal Domain Generalization? A Comprehensive Benchmark Study (NeurIPS 2026 · similarity 0.84)
- Smoothing the Shift: Towards Stable Test-Time Adaptation under Complex Multimodal Noises (ICLR 2025 · similarity 0.84)
- Extremely Simple Multimodal Outlier Synthesis for Out-of-Distribution Detection and Segmentation (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
