Less is more: Sampling chemical space with active learning
Open accessAbstract
The development of accurate and transferable machine learning (ML) potentials for predicting molecular energetics is a challenging task. The process of data generation to train such ML potentials is a task neither well understood nor researched in detail. In this work, we present a fully automated approach for the generation of datasets with the intent of training universal ML potentials. It is based on the concept of active learning (AL) via Query by Committee (QBC), which uses the disagreement between an ensemble of ML potentials to infer the reliability of the ensemble’s prediction. QBC allows the presented AL algorithm to automatically sample regions of chemical space where the ML potential fails to accurately predict the potential energy. AL improves the overall fitness of ANAKIN-ME (ANI) deep learning potentials in rigorous test cases by mitigating human biases in deciding what new training data to use. AL also reduces the training set size to a fraction of the data required when using naive random sampling techniques. To provide validation of our AL approach, we develop the COmprehensive Machine-learning Potential (COMP6) benchmark (publicly available on GitHub) which contains a diverse set of organic molecules. Active learning-based ANI potentials outperform the original random sampled ANI-1 potential with only 10% of the data, while the final active learning-based model vastly outperforms ANI-1 on the COMP6 benchmark after training to only 25% of the data. Finally, we show that our proposed AL technique develops a universal ANI potential (ANI-1x) that provides accurate energy and force predictions on the entire COMP6 benchmark. This universal ML potential achieves a level of accuracy on par with the best ML potentials for single molecules or materials, while remaining applicable to the general class of organic molecules composed of the elements CHNO.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
COMP6 benchmark was developed to validate the transferability of ML potentials.
“The COMP6 benchmark dataset191 was developed to validate the transferability of ML potentials.”
From QSAR without borders (2020) · cites this paper as [191]
COMP6 contains five rigorous benchmarks covering broad regions of organic and bio-chemical space of isolated molecules and a sixth built from the existing S66x8 noncovalent and intermolecular interactions data.
“COMP6 is a benchmark suite containing five rigorous benchmarks that cover broad regions of organic and bio-chemical space of isolated molecules and a sixth built from the existing S66x8195 noncovalent and intermolecular interactions data.191”
From QSAR without borders (2020) · cites this paper as [191]
Active learning methods provide consistent and automated improvement in accuracy and transferability of ML potentials and have contributed greatly to the success of general-purpose models.
“Active learning methods provide a consistent and automated improvement in accuracy and transferability and have contributed greatly to the success of general-purpose models.”
From QSAR without borders (2020) · cites this paper as [191]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
Rupp, M., Tkatchenko, A., Müller, K.-R., & von Lilienfeld, O. A. (2012) · Physical Review Letters · doi:10.1103/physrevlett.108.058301 · Cited by 2,312 · see citing works
Quantum chemistry structures and properties of 134 kilo molecules
Ramakrishnan, R., Dral, P. O., Rupp, M., & von Lilienfeld, O. A. (2014) · Scientific Data · doi:10.1038/sdata.2014.22 · Cited by 1,966 · see citing works
Quantum-chemical insights from deep tensor neural networks
Schütt, K. T., Arbabzadah, F., Chmiela, S., Müller, K. R., & Tkatchenko, A. (2017) · Nature Communications · doi:10.1038/ncomms13890 · Cited by 1,445 · see citing works
King, R. D., Rowland, J., Oliver, S. G., Young, M., Aubrey, W., Byrne, E., Liakata, M., Markham, M., Pir, P., Soldatova, L. N., et al. (2009) · Science · doi:10.1126/science.1165620 · Cited by 679 · see citing works
Brauer, B., Kesharwani, M. K., Kozuch, S., & Martin, J. M. L. (2016) · Physical Chemistry Chemical Physics · doi:10.1039/c6cp00688d · Cited by 238 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Closed-loop optimization of fast-charging protocols for batteries with machine learning
Attia, P. M., Grover, A., Jin, N., Severson, K. A., Markov, T. M., Liao, Y.-H., Chen, M. H., Cheong, B., Perkins, N., Yang, Z., et al. (2020) · Nature · doi:10.1038/s41586-020-1994-5 · Cited by 1,180 · see citing works
A Survey of Deep Active Learning
Ren, P., Xiao, Y., Chang, X., Huang, P.-Y., Li, Z., Gupta, B. B., Chen, X., & Wang, X. (2022) · ACM Computing Surveys · doi:10.1145/3472291 · Cited by 1,169 · see citing works
Muratov, E. N., Bajorath, J., Sheridan, R. P., Tetko, I. V., Filimonov, D., Poroikov, V., Oprea, T. I., Baskin, I. I., Varnek, A., Roitberg, A., et al. (2020) · Chemical Society Reviews · doi:10.1039/d0cs00098a · Cited by 913 · see citing works
Self-citationFrom DFT to machine learning: recent approaches to materials science–a review
Schleder, G. R., Padilha, A. C. M., Acosta, C. M., Costa, M., & Fazzio, A. (2019) · Journal of Physics: Materials · doi:10.1088/2515-7639/ab084b · Cited by 852 · see citing works
Mater, A. C., & Coote, M. L. (2019) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.9b00266 · Cited by 565 · see citing works
Opportunities and Challenges for Machine Learning in Materials Science
Morgan, D., & Jacobs, R. (2020) · Annual Review of Materials Research · doi:10.1146/annurev-matsci-070218-010015 · Cited by 549 · see citing works
Emerging materials intelligence ecosystems propelled by machine learning
Batra, R., Song, L., & Ramprasad, R. (2020) · Nature Reviews Materials · doi:10.1038/s41578-020-00255-y · Cited by 446 · see citing works
A quantitative uncertainty metric controls error in neural network-driven chemical discovery
Janet, J. P., Duan, C., Yang, T., Nandy, A., & Kulik, H. J. (2019) · Chemical Science · doi:10.1039/c9sc02298h · Cited by 243 · see citing works
Deep Learning in Protein Structural Modeling and Design
Gao, W., Mahajan, S. P., Sulam, J., & Gray, J. J. (2020) · Patterns · doi:10.1016/j.patter.2020.100142 · Cited by 227 · see citing works
A critical overview of computational approaches employed for COVID-19 drug discovery
Muratov, E. N., Amaro, R., Andrade, C. H., Brown, N., Ekins, S., Fourches, D., Isayev, O., Kozakov, D., Medina-Franco, J. L., Merz, K. M., et al. (2021) · Chemical Society Reviews · doi:10.1039/d0cs01065k · Cited by 206 · see citing works
Self-citationJanet, J. P., Ramesh, S., Duan, C., & Kulik, H. J. (2020) · ACS Central Science · doi:10.1021/acscentsci.0c00026 · Cited by 200 · see citing works
Strategies and Software for Machine Learning Accelerated Discovery in Transition Metal Chemistry
Nandy, A., Duan, C., Janet, J. P., Gugler, S., & Kulik, H. J. (2018) · Industrial & Engineering Chemistry Research · doi:10.1021/acs.iecr.8b04015 · Cited by 161 · see citing works
DPA-2: a large atomic model as a multi-task learner
Zhang, D., Liu, X., Zhang, X., Zhang, C., Cai, C., Bi, H., Du, Y., Qin, X., Peng, A., Huang, J., et al. (2023) · arXiv · doi:10.48550/arxiv.2312.15492 · Cited by 155
Vandermause, J., Xie, Y., Lim, J. S., Owen, C. J., & Kozinsky, B. (2022) · Nature Communications · doi:10.1038/s41467-022-32294-0 · Cited by 147 · see citing works
Neural scaling of deep chemical models
Frey, N. C., Soklaski, R., Axelrod, S., Samsi, S., Gómez-Bombarelli, R., Coley, C. W., & Gadepally, V. (2023) · Nature Machine Intelligence · doi:10.1038/s42256-023-00740-3 · Cited by 125 · see citing works
Active meta-learning for predicting and selecting perovskite crystallization experiments
Shekar, V., Nicholas, G., Najeeb, M. A., Zeile, M., Yu, V., Wang, X., Slack, D., Li, Z., Nega, P. W., Chan, E. M., et al. (2022) · The Journal of Chemical Physics · doi:10.1063/5.0076636 · Cited by 25 · see citing works
- Authors
- Smith, J. S., Nebgen, B., Lubbers, N., Isayev, O., & Roitberg, A. E.
- Affiliations
- University of Florida (US)
- Los Alamos National Laboratory 2 , Los Alamos, New Mexico 87545, USA
- Los Alamos National Laboratory (US)
- University of North Carolina at Chapel Hill (US)
- Department of Chemistry, University of Florida 1 , Gainesville, Florida 32611, USA
- Corresponding author
- Isayev, Olexandr
- Venue
- The Journal of Chemical Physics
- Publisher
- AIP Publishing
- Published
- 2018-05-22
- Volume
- 148
- Issue
- 24
- Pages
- 241733-241733
- ISSN
- 0021-9606 (ISSN-L)
- 1089-7690
- OpenAlex
- W2785813126
- Cited by
- 811
- References
- 5
- Citations by year
- 2026: 89 · 2025: 138 · 2024: 123 · 2023: 110 · 2022: 102 · 2021: 93 · 2020: 94 · 2019: 49 · 2018: 13
- FWCI
- 23.24
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- The Journal of Chemical Physics (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Division of Physics (1148698)
- U.S. Department of Energy (DE-AC52-06NA25396)
- Office of Naval Research (N00014-16-1-2311)
- Division of Advanced Cyberinfrastructure (ACI-1053575)
- Topics
- Physical Sciences › Materials Science › Materials Chemistry › Machine Learning in Materials Science
- Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Advanced Chemical Physics Studies
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Computational Drug Discovery Methods
- Keywords
- active learning (machine learning), query-by-committee, machine learning interatomic potentials, training data generation, organic compounds, molecular energetics, force prediction, transferable potentials
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
