2020/10/14 by Yuan Tian, Dezhen Xue, Ruihao Yuan +4 · 25 citations
Computer Science · Materials Science · Physics and Astronomy · #Active learning (machine learning) #Curve fitting #Data point #Function (biology) #Gaussian Processes and Bayesian Inference #Liquidus #Machine Learning in Materials Science #Model Reduction and Neural Networks #Property (philosophy) #Sampling (signal processing) #Set (abstract data type) #Surface (topology) #cond-mat.mtrl-sci #physics.comp-ph
paper · pdf · doi:10.1103/physrevmaterials.5.013802
published in Physical Review Materials 5(1) (American Physical Society) · 15 pages and 15 figures,submitted to Physical Review Materials
arxiv created 2020/10/14 · openalex created_date 2020/10/22 · openalex publication_date 2021/01/08 · arxiv updated 2021/01/13 · openalex updated_date 2026/08/05
The relationship between material properties and independent variables such as temperature, external field, or time is usually represented by a curve or surface in a multidimensional space. Determining such a curve or surface requires a series of experiments or calculations which are often time and cost consuming. A general strategy uses an appropriate utility function to sample the space to recommend the next optimal experiment or calculation within an active learning loop. However, knowing what optimal sampling strategy to use to minimize the number of experiments is an outstanding problem. We compare a number of strategies based on directed exploration on several materials problems of varying complexity using a Kriging-based model. These include one-dimensional curves such as the fatigue life curve for 304L stainless steel and the Liquidus line of the Fe-C phase diagram, surfaces such as the Hartmann 3 function in three-dimensional space and the fitted intermolecular potential for Ar-SH, and a four-dimensional data set of experimental measurements for BaTiO3-based ceramics. We also consider the effects of experimental noise on the Hartmann 3 function. We find that directed exploration guided by maximum variance provides better performance overall, converging faster across several data sets. However, for certain problems, the tradeoff methods incorporating exploitation can perform at least as well, if not better than maximum variance. Thus, we discuss how the choice of the utility function depends on the distribution of the data, the model performance and uncertainties, additive noise, as well as the budget.