AI Paper Atlas

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

NeurIPS 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Prototype and concept dictionary models · 6 citations (OpenAlex)

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 changedWhereMeasuredEffect
add introspective stage H on ResNet-18 featuresResNet-18 · CIFAR-10Caccuracy (%): 67.89 -> 71.4Improved
add introspection on top of DenoisingDenoising · CIFAR-10Caccuracy (%): 65.02 -> 68.86Improved
add introspection on top of Adversarial TrainAdversarial Train · CIFAR-10Caccuracy (%): 68.02 -> 70.86Improved
add introspection on top of SimCLRSimCLR · CIFAR-10Caccuracy (%): 70.28 -> 73.32Improved
add introspection on top of Augment NoiseAugment Noise · CIFAR-10Caccuracy (%): 76.86 -> 77.98Improved
add introspection on top of AugmixAugmix WideResNet · CIFAR-10Caccuracy (%): 89.85 -> 89.89No clear effect
add introspection on top of AugmixAugmix WideResNet · CIFAR-10CECE: 43.33Improved
introspection when train and test distributions matchResNet-18 · CIFAR-10 test setaccuracy (%): 91.02 -> 90.93No clear effect
introspection when train and test distributions matchResNet-50 · CIFAR-10 test setaccuracy (%): 93.09 -> 93.17No clear effect
sample active learning queries with H instead of f (Entropy)Entropy strategy ResNet-18 · CIFAR-10C Gaussian noiseaccuracy: 0.244 -> 0.258Improved
sample active learning queries with H instead of f (BALD)BALD ResNet-18 · CIFAR-10C Gaussian noiseaccuracy: 0.26 -> 0.273Improved
sample active learning queries with H on original testsetEntropy ResNet-18 · CIFAR-10 original testsetaccuracy: 0.365 -> 0.365No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy