Research map / Label-efficient and robust learning / Self-supervised and contrastive
"Trust, but Verify: Using Self-Supervised Probing to Improve Trustworthiness"
Ailin Deng, Shen Li, Miao Xiong, Zhirui Chen, Bryan Hooi
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 self-supervised probing confidence to MSP | MSP · CIFAR-10 misclassification (ResNet-18) | AUROC: 90.63 -> 92.10 | Improved |
| add self-supervised probing to MCDropout | MCDropout · CIFAR-10 misclassification (ResNet-18) | AUPR-ERR: 42.04 -> 52.60 | Improved |
| add self-supervised probing to TrustScore | TrustScore · CIFAR-10 misclassification (VGG16) | AUROC: 90.47 -> 91.68 | Improved |
| add self-supervised probing to MSP | MSP · CINIC-10 misclassification (VGG16) | AUROC: 86.33 -> 85.53 | Hurt |
| add self-supervised probing to TCP | TCP · CIFAR-10 misclassification (ResNet-18) | AUROC: 93.74 -> 93.73 | No clear effect |
| add translation probing task to rotation probing (ablation) | SSP · CIFAR-10 misclassification (ResNet-18) | AUROC: 91.78 -> 92.10 | Improved |
| add rotation probing task to translation probing (ablation) | SSP · CIFAR-10 misclassification (ResNet-18) | AUROC: 91.42 -> 92.10 | Improved |
| add SSP to MSP for OOD detection | MSP · CIFAR-10 to SVHN OOD (VGG16) | AUROC: 91.05 -> 92.57 | Improved |
| add SSP to Entropy for OOD detection | Entropy · CIFAR-10 to LSUN OOD (ResNet-18) | AUROC: 91.94 -> 93.55 | Improved |
| add SSP to MSP for OOD detection | MSP · CINIC-10 to SVHN OOD (VGG16) | AUROC: 81.48 -> 85.25 | Improved |
| add input-dependent temperature from probing to temperature scaling | Temp. Scaling · CIFAR-10 calibration (ResNet-18) | ECE: 1.40 -> 0.75 | Improved |
| add input-dependent temperature from probing to temperature scaling | Temp. Scaling · CIFAR-10 calibration (VGG16) | MCE: 7.53 -> 9.15 | Hurt |
Closest papers by meaning
- Towards Trustworthy Deep Learning (AAAI 2024 · similarity 0.85)
- Trustworthy Machine Learning through Data-Specific Indistinguishability (ICML 2025 · similarity 0.85)
- U-trustworthy Models. Reliability, Competence, and Confidence in Decision-Making (AAAI 2024 · similarity 0.84)
- The Mainstays of Trustworthy Machine Learning (AAAI 2025 · similarity 0.84)
- Decoding Compressed Trust: Scrutinizing the Trustworthiness of Efficient LLMs Under Compression (ICML 2024 · similarity 0.83)
- CURE4Rec: A Benchmark for Recommendation Unlearning with Deeper Influence (NeurIPS 2024 · similarity 0.83)
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.
