Efficient learning of accurate surrogates for simulations of complex systems

Journal Article · 17 May 2024 · Nature Machine Intelligence

DOI: https://doi.org/10.1038/s42256-024-00839-1 · OSTI ID:2426829

Diaw, A.  ^[1]; McKerns, M.  ^[2]; Sagert, I.  ^[2]; Stanton, L. G. ^[3]; Murillo, M. S. ^[4]

---

1. Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
2. Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
3. San Jose State Univ., CA (United States)
4. Michigan State Univ., East Lansing, MI (United States)

Machine learning methods are increasingly deployed to construct surrogate models for complex physical systems at a reduced computational cost. However, the predictive capability of these surrogates degrades in the presence of noisy, sparse or dynamic data. Here, we introduce an online learning method empowered by optimizer-driven sampling that has two advantages over current approaches: it ensures that all local extrema (including endpoints) of the model response surface are included in the training data, and it employs a continuous validation and update process in which surrogates undergo retraining when their performance falls below a validity threshold. We find, using benchmark functions, that optimizer-directed sampling generally outperforms traditional sampling methods in terms of accuracy around local extrema even when the scoring metric is biased towards assessing overall accuracy. Finally, the application to dense nuclear matter demonstrates that highly accurate surrogates for a nuclear equation-of-state model can be reliably autogenerated from expensive calculations using few model evaluations.

• Research Organization: Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
• Sponsoring Organization: USDOE Laboratory Directed Research and Development (LDRD) Program; USDOE National Nuclear Security Administration (NNSA); National Science Foundation (NSF)
• Grant/Contract Number: 89233218CNA000001
• OSTI ID: 2426829
• Report Number(s): LA-UR--20-24947
• Journal Information: Nature Machine Intelligence, Journal Name: Nature Machine Intelligence Journal Issue: 5 Vol. 6; ISSN 2522-5839
• Publisher: Springer NatureCopyright Statement
• Country of Publication: United States
• Language: English