AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

DISCO: Diversifying Sample Condensation for Accelerating Model Evaluation

Alexander Rubinstein, Benjamin Raible, Martin Gubri, Seong Joon Oh

ICLR 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Active learning and subset selection

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 model signatures with Random Forest on uniform random samples (ablation)direct evaluation on random samples · MMLUMAE and rank: 3.45 -> 1.81 MAE; .916 -> .933 rankImproved
select anchor points with high PDS instead of uniform random (ablation)Sig. + RF · MMLUMAE and rank: 1.81 -> 1.07 MAE; .933 -> .987 rankImproved
replace chronological split by uniform split (ablation)DISCO · MMLUrank: .987 vs .986No clear effect
add task stratification of anchors (ablation)DISCO with PDS · MMLUrank: .978 vs .987No clear effect
reduce source models to 100 from 382 (ablation)DISCO · MMLUrank: .987 -> .969Hurt
add PCA dimensionality reduction to 256 dims (ablation)DISCO · MMLUrank: .918 -> .987Improved
use Random Forest instead of other prediction models (ablation)DISCO · MMLUrank: .987Improved
use kNN instead of Random Forest at 10 samples (ablation)DISCO · MMLUMAE and rankImproved
add PDS selection with Random Forest on vision (ablation)Sig. + RF · ImageNet-1k valMAE and rank: 0.86 -> 0.63 MAE; .944 -> .969 rankImproved
add model signatures with RF on uniform samples (vision) (ablation)direct evaluation · ImageNet-1k valMAE and rank: 3.03 -> 0.86 MAE; 0.652 -> 0.944 rankImproved
use High JSD instead of High PDS with kNN (ablation)DISCO · HellaSwagMAE: 1.32 -> 1.50 MAEHurt

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