mlsanityData to 7 Oct 2026

A Critical Examination of Active Learning Workflows in Materials Science

arXiv 2601.05946 ↗First posted 9 Jan 2026cond-mat.mtrl-sci cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Active learning (AL) plays a critical role in materials science, enabling applications such as the construction of machine-learning interatomic potentials for atomistic simulations and the operation of self-driving laboratories. Despite its widespread use, the reliability and effectiveness of AL workflows depend on implicit design assumptions that are rarely examined systematically. Here, we critically assess AL workflows deployed in materials science and investigate how key design choices, such as surrogate models, sampling strategies, uncertainty quantification and evaluation metrics, relate to their performance. By identifying common pitfalls and discussing practical mitigation strategies, we provide guidance to practitioners for the efficient design, assessment, and interpretation of AL workflows in materials science.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 28 of its 128 references that are in the map. First posted on arXiv 9 Jan 2026.

Builds on

28 of 128 references.

References · 28

  1. 2017Learning Active Learning from Datacited by 100
  2. 2018Less is more: sampling chemical space with active learningcited by 80
  3. 2021On Statistical Bias In Active Learning: How and When To Fix Itcited by 72
  4. 2021Uncertainty Toolbox: an Open-Source Library for Assessing, Visualizing, and Improving Uncertainty Quantificationcited by 57
  5. 2021Benchmarking the Performance of Bayesian Optimization across Multiple Experimental Materials Science Domainscited by 46
  6. 2018Active Learning of Uniformly Accurate Inter-atomic Potentials for Materials Simulationcited by 37
  7. 2020Olympus: a benchmarking framework for noisy optimization and experiment planningcited by 28
  8. 2022In Defense of Core-set: A Density-aware Core-set Selection for Active Learningcited by 26
  9. 2019Exploring Representativeness and Informativeness for Active Learningcited by 21
  10. 2022Data-driven discovery of novel 2D materials by deep generative modelscited by 18
  11. 2024Structure-based out-of-distribution (OOD) materials property prediction: a benchmark studycited by 17
  12. 2022Materials Property Prediction with Uncertainty Quantification: A Benchmark Studycited by 14
16 more
  1. 2023Calibration in Machine Learning Uncertainty Quantification: beyond consistency to target adaptivitycited by 13
  2. 2024Model-free quantification of completeness, uncertainties, and outliers in atomistic machine learning using information theorycited by 10
  3. 2022Benchmarking Active Learning Strategies for Materials Optimization and Discoverycited by 9
  4. 2023Clarifying Trust of Materials Property Predictions using Neural Networks with Distribution-Specific Uncertainty Quantificationcited by 8
  5. 2024Best Practices for Multi-Fidelity Bayesian Optimization in Materials and Molecular Researchcited by 7
  6. 2019Machine-learned metrics for predicting the likelihood of success in materials discoverycited by 6
  7. 2022PI is back! Switching Acquisition Functions in Bayesian Optimizationcited by 6
  8. 2022ET-AL: Entropy-Targeted Active Learning for Bias Mitigation in Materials Datacited by 6
  9. 2024A General Approach for Determining Applicability Domain of Machine Learning Modelscited by 5
  10. 2021Addressing Bias in Active Learning with Depth Uncertainty Networks... or Notcited by 3
  11. 2023A Data Driven Sequential Learning Framework to Accelerate and Optimize Multi-Objective Manufacturing Decisionscited by 3
  12. 2023Overcoming Overconfidence for Active Learningcited by 3
  13. 2024Improved Uncertainty Estimation of Graph Neural Network Potentials Using Engineered Latent Space Distancescited by 3
  14. 2024Batched Bayesian optimization by maximizing the probability of including the optimumcited by 2
  15. 2024Accelerating the Training and Improving the Reliability of Machine-Learned Interatomic Potentials for Strongly Anharmonic Materials through Active Learningcited by 1
  16. 2025Steering an Active Learning Workflow Towards Novel Materials Discovery via Queue Prioritizationcited by 1

Built on by

Not cited yet.