Efficient Learning of Accurate Surrogates for Simulations of Complex Systems
arXiv:2207.12855 · doi:10.1038/s42256-024-00839-1
Abstract
Machine learning methods are increasingly used to build computationally inexpensive surrogates for complex physical models. The predictive capability of these surrogates suffers when data are noisy, sparse, or time-dependent. As we are interested in finding a surrogate that provides valid predictions of any potential future model evaluations, we introduce an online learning method empowered by optimizer-driven sampling. The method has two advantages over current approaches. First, it ensures that all turning points on the model response surface are included in the training data. Second, after any new model evaluations, surrogates are tested and "retrained" (updated) if the "score" drops below a validity threshold. Tests on benchmark functions reveal that optimizer-directed sampling generally outperforms traditional sampling methods in terms of accuracy around local extrema, even when the scoring metric favors overall accuracy. We apply our method to simulations of nuclear matter to demonstrate that highly accurate surrogates for the nuclear equation of state can be reliably auto-generated from expensive calculations using a few model evaluations.
arXiv comment13 pages, 6 figures, submitted to Nature Machine Intelligence
References in corpus (19)
- GW170817: Observation of Gravitational Waves from a Binary Neutron Star Inspiral
- PSR J0030+0451 Mass and Radius from NICER Data and Implications for the Properties of Neutron Star Matter
- A NICER View of PSR J0030+0451: Millisecond Pulsar Parameter Estimation
- Heavy Ion Collisions: The Big Picture, and the Big Questions
- From hadrons to quarks in neutron stars: a review
- Evidence for quark-matter cores in massive neutron stars
- Stringent constraints on neutron-star radii from multimessenger observations and nuclear theory
- Properties of hot and dense matter from relativistic heavy ion collisions
- Quarkyonic Matter and Neutron Stars
- Path Integral Monte Carlo Simulation of the Warm-Dense Homogeneous Electron Gas
- Quark phases in neutron stars and a "third family" of compact stars as a signature for phase transitions
- Non-monotonic energy dependence of net-proton number fluctuations
- Fast machine-learning online optimization of ultra-cold-atom experiments
- Core-collapse supernova explosions triggered by a quark-hadron phase transition during the early post-bounce phase
- Nuclear and neutron-star matter from local chiral interactions
- The APR equation of state for simulations of supernovae, neutron stars and binary mergers
- Finite-temperature extension for cold neutron star equations of state
- Conditions for Phase Equilibrium in Supernovae, Proto-Neutron and Neutron Stars
- Tabulated Neutron Star Equations of State Modeled within the Chiral Mean Field Model