Research map / Label-efficient and robust learning / Recognition and long-tail
Introspective Learning : A Two-Stage approach for Inference in Neural Networks
Mohit Prabhushankar, Ghassan AlRegib
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 introspective stage H on ResNet-18 features | ResNet-18 · CIFAR-10C | accuracy (%): 67.89 -> 71.4 | Improved |
| add introspection on top of Denoising | Denoising · CIFAR-10C | accuracy (%): 65.02 -> 68.86 | Improved |
| add introspection on top of Adversarial Train | Adversarial Train · CIFAR-10C | accuracy (%): 68.02 -> 70.86 | Improved |
| add introspection on top of SimCLR | SimCLR · CIFAR-10C | accuracy (%): 70.28 -> 73.32 | Improved |
| add introspection on top of Augment Noise | Augment Noise · CIFAR-10C | accuracy (%): 76.86 -> 77.98 | Improved |
| add introspection on top of Augmix | Augmix WideResNet · CIFAR-10C | accuracy (%): 89.85 -> 89.89 | No clear effect |
| add introspection on top of Augmix | Augmix WideResNet · CIFAR-10C | ECE: 43.33 | Improved |
| introspection when train and test distributions match | ResNet-18 · CIFAR-10 test set | accuracy (%): 91.02 -> 90.93 | No clear effect |
| introspection when train and test distributions match | ResNet-50 · CIFAR-10 test set | accuracy (%): 93.09 -> 93.17 | No clear effect |
| sample active learning queries with H instead of f (Entropy) | Entropy strategy ResNet-18 · CIFAR-10C Gaussian noise | accuracy: 0.244 -> 0.258 | Improved |
| sample active learning queries with H instead of f (BALD) | BALD ResNet-18 · CIFAR-10C Gaussian noise | accuracy: 0.26 -> 0.273 | Improved |
| sample active learning queries with H on original testset | Entropy ResNet-18 · CIFAR-10 original testset | accuracy: 0.365 -> 0.365 | No clear effect |
Closest papers by meaning
- Explainable Neural Networks with Guarantee: A Sparse Estimation Approach (AAAI 2025 · similarity 0.86)
- Latent Introspection: Models Can Detect Prior Concept Injections (NeurIPS 2026 · similarity 0.85)
- Do We Always Need the Simplicity Bias? Looking for Optimal Inductive Biases in the Wild (CVPR 2025 · similarity 0.84)
- Unlearning-based Neural Interpretations (ICLR 2025 · similarity 0.84)
- Neural Collapse in Cumulative Link Models for Ordinal Regression: An Analysis with Unconstrained Feature Model (NeurIPS 2025 · similarity 0.84)
- Bridging Critical Gaps in Convergent Learning: How Representational Alignment Evolves Across Layers, Training, and Distribution Shifts (NeurIPS 2025 · 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.
