AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

What shapes feature representations? Exploring datasets, architectures, and training

Katherine Hermann, Andrew Lampinen

NeurIPS 2020 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 62 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
Train on a target feature rather than leave the model untrainedAlexNet and ResNet-50 · Trifeature and Navontarget-feature decoding accuracyImproved
Train on a target feature, which suppresses unrelated featuresAlexNet and ResNet-50 · Trifeature and Navonnon-target feature decoding accuracyHurt
Prefer color over shape when both redundantly predict the labelAlexNet and ResNet-50 · Trifeature correlated pairsfeature decoding accuracyMixed
Prefer shape over texture when both redundantly predict the labelAlexNet and ResNet-50 · Trifeature correlated pairsfeature decoding accuracyMixed
Increase correlation of the non-target feature with the perfectly predictive targetAlexNet and ResNet-50 · Trifeature correlated datasetscorrelated non-target feature decoding accuracyNo clear effect
Use the easy linear feature as it becomes more predictive5-layer MLP · Binary Featuresfeature relianceHurt
Train two models on the same task rather than different tasksVision models · Trifeature and Navonrepresentational similarity: 0.408 [0.394, 0.420] versus 0.624 [0.615, 0.632]Improved
Train on texture rather than shapeVision models · Trifeaturerepresentational similarity: 0.408 [0.394, 0.420] versus 0.624 [0.615, 0.632]Hurt
Train on an easy feature rather than an unpredictive easy featureBinary-feature models · Binary Featuresrepresentational similarity across model runsImproved
Train a multitask model to report both easy and difficult featuresBinary-feature model · Binary Featuresrepresentational similarity to easy-feature modelNo 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