AI Paper Atlas

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

Deep Learning for BioImaging: What Are We Really Learning?

Ivan Svatko, Maxime Sanchez, Ihab Bendidi, Gilles Cottrell, Auguste Genovesio

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 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
pretrain the image encoder instead of leaving it untrainedfull-image encoders · HEST-1k-1NN CCRCCgene-wise PCC: 0.12 -> 0.37Improved
pretrain the image encoder instead of leaving it untrainedfull-image encoders · HEST-1k-1NN COADgene-wise PCC: 0.27 -> 0.38Improved
use in-domain foundation model instead of best structure-only modelimage vs structure · HEST-1k-1NN CCRCCgene-wise PCC: 0.16 -> 0.37Improved
use structure-only model instead of out-of-domain image encoderstructure-only encoder · HEST-1k-1NN COADgene-wise PCC: 0.31 vs 0.30No clear effect
pretrain GCN structure encoder (vs cell count baseline) (ablation)GCN · HEST-1k-1NN SKCMgene-wise PCC: 0.40 -> 0.45Improved
pretrain GCN structure encoder (vs cell count baseline) (ablation)GCN · HEST-1k-1NN CCRCCgene-wise PCC: 0.11 -> 0.04Hurt
use EGNN instead of GCN (ablation)structure-only encoder · HEST-1k-1NN SKCMgene-wise PCC: 0.45 -> 0.22Hurt
pretrain image encoders on RxRx3-core gene retrieval (ablation)ViT · RxRx3-core recall@5%recall@5%No clear effect
use in-domain MAE-L/8 and MAE-G/8 pretrainingMAE backbones · RxRx3-core recall@5%recall@5%: 1.5 times higher averaged top 5% recallImproved
use ResNet instead of ViTimage encoders · RxRx3-core recall@5%recall@5%Hurt
use deeper layers for representation (ablation)image encoders · RxRx3-core recall@5%recall@5%Hurt
pretrain on cell-painting data (in-domain) rather than ImageNetOpenPhenom vs OOD · JUMP-CP mAPmAPImproved

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