Active learning of molecular data for task-specific objectives

doi:10.1063/5.0229834

2025-01-02

DOI：10.1063/5.0229834

原文链接

原文求助

摘要

En 中文

Active learning (AL) has shown promise to be a particularly data-efficient machine learning approach. Yet, its performance depends on the application, and it is not clear when AL practitioners can expect computational savings. Here, we carry out a systematic AL performance assessment for three diverse molecular datasets and two common scientific tasks: compiling compact, informative datasets and targeted molecular searches. We implemented AL with Gaussian processes (GP) and used the many-body tensor as molecular representation. For the first task, we tested different data acquisition strategies, batch sizes, and GP noise settings. AL was insensitive to the acquisition batch size, and we observed the best AL performance for the acquisition strategy that combines uncertainty reduction with clustering to promote diversity. However, for optimal GP noise settings, AL did not outperform the randomized selection of data points. Conversely, for targeted searches, AL outperformed random sampling and achieved data savings of up to 64%. Our analysis provides insight into this task-specific performance difference in terms of target distributions and data collection strategies. We established that the performance of AL depends on the relative distribution of the target molecules in comparison to the total dataset distribution, with the largest computational savings achieved when their overlap is minimal.

Keyword:

ORBITAL ENERGIES

PREDICTIONS

POTENTIALS

期刊

Journal of Chemical Physics

IF:

3.1

论文数:

7.2W

被引数:

23.2W

机构

Aalto University

学者数:

1.6W

论文数: 1.5W

被引数: 2.1W

University of Turku

学者数:

1.7W

论文数: 1.5W

被引数: 2.0W

引用论文

Geometric deep learning on molecular representations 分子表征的几何深度学习

NATURE MACHINE INTELLIGENCE

IF23.9

2021-12-15

160

PREAI

Atz, Kenneth; Grisoni, Francesca; Schneider, Gisbert

ACTIVE LEARNING TO OVERCOME SAMPLE SELECTION BIAS: APPLICATION TO PHOTOMETRIC VARIABLE STAR CLASSIFICATION

ASTROPHYSICAL JOURNAL

IF5.4

2011-12-22

OAAI

Richards, Joseph W.; Starr, Dan L.; Brink, Henrik; Miller, Adam A.; Bloom, Joshua S.; Butler, Nathaniel R.; James, J. Berian; Long, James P.; Rice, John

An introduction to ROC analysis ROC分析导论

PATTERN RECOGNITION LETTERS

IF3.3

2006-06-01

1.4W

PREAI

Fawcett, Tom

Machine learning potentials for metal-organic frameworks using an incremental learning approach

NPJ COMPUTATIONAL MATERIALS

IF11.9

2023-02-06

OAAI