AI Paper Atlas

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

Reducing Data Complexity Using Autoencoders With Class-Informed Loss Functions

David Charte, Francisco Charte, Francisco Herrera

TPAMI 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 22 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 class-informed Slicer penalty to the autoencoder (ablation)AE · MLP classifier F-scoreaverage rank (lower is better): 4.731 -> 2.000Improved
add class-informed Slicer penalty to the autoencoder (ablation)AE · kNN classifier F-scoreaverage rank (lower is better): 4.115 -> 2.212Improved
add class-informed Slicer penalty to the autoencoder (ablation)AE · SVM classifier F-scoreaverage rank (lower is better): 3.538 -> 2.692Improved
add KLD-based Skaler penalty to the autoencoder (ablation)AE · MLP classifier F-scoreaverage rank (lower is better): 4.731 -> 2.596Improved
add Fisher-ratio Scorer penalty to the autoencoder (ablation)AE · MLP classifier F-scoreaverage rank (lower is better): 4.731 -> 3.865Improved
add Fisher-ratio Scorer penalty to the autoencoder (ablation)AE · SVM classifier F-scoreaverage rank (lower is better): 3.538 -> 2.923Improved
add KLD-based Skaler penalty to the autoencoder (ablation)AE · complexity metric F1average rank (lower is better): 4.038 -> 1.577Improved
add Slicer penalty to the autoencoder (ablation)AE · complexity metric F3average rank (lower is better): 5.654 -> 2.231Improved
use Slicer instead of PCA (ablation)PCA · kNN classifier F-scoreaverage rank (lower is better): 3.096 -> 2.212Improved
use Slicer instead of PCA (ablation)PCA · complexity metric L2average rank (lower is better): 2.712 -> 2.846Hurt
use Skaler (F1-separability winner) instead of Slicer (ablation)Slicer · complexity metric F1 vs classifieraverage rank (lower is better): 1.577 vs 2.346 on F1; 2.596 vs 2.000 on MLP F-scoreMixed
use Slicer instead of IsomapIsomap · kNN classifier F-scoreaverage rank (lower is better): 4.904 -> 2.212Improved

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