AI Paper Atlas

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

Learning Coherent Representations: A Topological Approach to Interpretability

Sigurd Gaukstad, Melvin Vaupel, Valdemar Kargård Olsen, Erik Hermansen, Benjamin Adric Dunn

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Self-explaining and visual explanation networks · 0 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 COH coherence lossautoencoder · single rotated digit 6tuned features % (MRL>0.5): 63.3 -> 100.0 (vs L1)Improved
add COH coherence lossautoencoder · single rotated digit 6MSE: 0.010 -> 0.011 (L1 vs COH)No clear effect
add COH coherence loss instead of L1 sparsityautoencoder · single rotated digit 6sparsity (active % of features): 19.5 -> 33.2Mixed
add COH coherence lossautoencoder · two circles toypure features %: 1.2 -> 90.2 (vs L1)Improved
add COH coherence lossautoencoder · two circles toylocality (coherence epsilon): 4.62 -> 0.14 (L1 vs COH)Improved
add L1 sparsity alone (ablation)autoencoder · two circles toylocality: 0.53 -> 4.62 (vanilla vs L1)Hurt
add COH coherence lossautoencoder · two rotated digits 3 and 7MRL and tuned scoreMixed
add L1 sparsity to COH (ablation)autoencoder · two rotated digits 3 and 7class separation and angle tuning across seedsImproved
add COH loss to non-negative Softplus token embedding (ablation)BERT token embedding · WikiText-2features with >50% overlap (of 256): 1.0 -> 77.4Improved
add COH loss to non-negative Softplus token embedding (ablation)BERT token embedding · WikiText-2Claude-judged interpretable features (of 256): 0.0 -> 87.6Improved
add COH loss to non-negative Softplus token embedding (ablation)BERT token embedding · WikiText-2mean overlap: 0.22 -> 0.45Improved
add COH lossBERT token embedding · WikiText-2model performanceNo 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