AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

FUSION: Dataset Pruning via Fusing Uncertainty with Structural Information for Optimal Neural Training in Crystal Property Prediction

Xiean Wang, Pin Chen, Liqin Tan, Yutong Lu, Qingsong Zou

AAAI 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Machine-learned interatomic potentials · 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
use FUSION (uncertainty + structure) instead of random pruningALIGNN · Perovskites OOD, 80% prunedMAE: 0.223 -> 0.127Improved
use FUSION instead of random pruningALIGNN · Dielectric OOD, 80% prunedMAE: 0.421 -> 0.319Improved
use FUSION instead of random pruning across 11 architectures11 models · Perovskites (ID)MAE improvement: 13.65% averageImproved
use FUSION instead of random pruning across 11 architectures11 models · Dielectric (ID)MAE improvement: 1.91% averageImproved
use FUSION instead of diversity samplingALIGNN · JDFT2D ID, 80% prunedMAE: 12.7% MAE reductionImproved
add structural weighting factor to uncertainty (full FUSION vs uncertainty only) (ablation)FUSION · Dielectric ID, 80% prunedMAE: 0.940 vs FUSION (161% worse)Improved
add uncertainty to structural weighting factor (full FUSION vs weighting only) (ablation)FUSION · Perovskites OOD, 80% prunedMAE: 0.200 -> 0.127Improved
move from ID to OOD evaluationFUSION vs second-best · JDFT2D 80% prunedMAE gap: 9.752 -> 0.185 MAE gapHurt
prune a small optimal fraction (0-8.5%) of training datavarious models · Matbench datasetsMAE: up to 12.67% MAE reductionImproved
use ensemble-disagreement active learning for pruningALIGNN · JDFT2D ID, 80% prunedMAE: 131.210Hurt
FUSION vs random pruning at 40% pruningALIGNN · Perovskites OOD, 40% prunedMAE: 0.098 -> 0.100No 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