A framework for materials informatics education through workshops.

Abstract: The burgeoning field of materials informatics necessitates a focus on educating the next generation of materials scientists in the concepts of data science, artificial intelligence (AI), and machine learning (ML). In addition to incorporating these topics in undergraduate and graduate curricula, regular hands-on workshops present the most effective medium to initiate researchers to informatics and have them start applying the best AI/ML tools to their own research. With the help of the Materials Research Society (MRS), members of the MRS AI Staging Committee, and a dedicated team of instructors, we successfully conducted workshops covering the essential concepts of AI/ML as applied to materials data, at both the Spring and Fall Meetings in 2022, with plans to make this a regular feature in future meetings. In this article, we discuss the importance of materials informatics education via the lens of these workshops, including details such as learning and implementing specific algorithms, the crucial nuts and bolts of ML, and using competitions to increase interest and participation.

Leveraging uncertainty estimates and derivative information in Gaussian process regression for efficient collection and use of molecular simulation data.

We introduce Gaussian Process Regression (GPR) as an enhanced method of thermodynamic extrapolation and interpolation. The heteroscedastic GPR models that we introduce automatically weight provided information by its estimated uncertainty, allowing for the incorporation of highly uncertain, high-order derivative information. By the linearity of the derivative operator, GPR models naturally handle derivative information and, with appropriate likelihood models that incorporate heterogeneous uncertainties, are able to identify estimates of functions for which the provided observations and derivatives are inconsistent due to the sampling bias that is common in molecular simulations. Since we utilize kernels that form complete bases on the function space to be learned, the estimated uncertainty in the model takes into account that of the functional form itself, in contrast to polynomial interpolation, which explicitly assumes the functional form to be fixed. We apply GPR models to a variety of data sources and assess various active learning strategies, identifying when specific options will be most useful. Our active-learning data collection based on GPR models incorporating derivative information is finally applied to tracing vapor-liquid equilibrium for a single-component Lennard-Jones fluid, which we show represents a powerful generalization to previous extrapolation strategies and Gibbs-Duhem integration. A suite of tools implementing these methods is provided at https://github.com/usnistgov/thermo-extrap.

Leveraging Theory for Enhanced Machine Learning.

The application of machine learning to the materials domain has traditionally struggled with two major challenges: a lack of large, curated data sets and the need to understand the physics behind the machine-learning prediction. The former problem is particularly acute in the polymers domain. Here we aim to simultaneously tackle these challenges through the incorporation of scientific knowledge, thus, providing improved predictions for smaller data sets, both under interpolation and extrapolation, and a degree of explainability. We focus on imperfect theories, as they are often readily available and easier to interpret. Using a system of a polymer in different solvent qualities, we explore numerous methods for incorporating theory into machine learning using different machine-learning models, including Gaussian process regression. Ultimately, we find that encoding the functional form of the theory performs best followed by an encoding of the numeric values of the theory.

Graphene and Poly(3,4-ethylene dioxythiophene):Poly(4-styrenesulfonate) on Nonwoven Fabric as a Room Temperature Metal and Its Application as Dry Electrodes for Electrocardiography.

Highly conductive, metal-like poly(ethylene terephthalate) (PET) nonwoven fabric was prepared by coating poly(3,4-ethylenedioxythiophene):poly(4-styrenesulfonate) (PEDOT:PSS) containing dimethyl sulfoxide (DMSO) onto PET nonwoven fabric previously coated with graphene/graphite. The sheet resistance of the original nonwoven fabric decreases from >80 MΩ□-1 to 1.1 Ω□-1 after coating with 10.7 wt % graphene and 5.48 wt % PEDOT:PSS with a maximum current at breakdown of 4 A. This sheet resistance is lower than previously reported sheet resistances of fabrics coated with graphene films, PEDOT:PSS films, or PEDOT:PSS coated fabrics from the literature. The effect of temperature on the resistance of graphene/PEDOT:PSS coated fabric has revealed that the resistance decreases with increasing temperature, analogous to semiconductors, with a clear semiconductor-metal transition occurring at 290 K. Finally, a coating of 18 wt % graphene/graphite and 2.5 wt % PEDOT:PSS (Rs = 5.5 Ω□-1) screen printed on the nonwoven fabric was shown to function as an electrode for electrocardiography without any hydrogel and with dry skin conditions. This composite coating finds application in wearable electronics for military and consumer sectors.

Switchable 3-0 magnetoelectric nanocomposite thin film with high coupling.

A mixed precursor solution method was used to deposit 3-0 nanocomposite thin films of PbZr0.52Ti0.48O3 (PZT) and CoFe2O4 (CFO). The piezoelectric behavior of PZT and magnetostrictive behavior of CFO allow for magnetoelectric (ME) coupling through strain transfer between the respective phases. High ME coupling is desired for many applications including memory devices, magnetic field sensors, and energy harvesters. The spontaneous phase separation in the 3-0 nanocomposite film was observed, with 25 nm CFO particle or nanophases distributed in discrete layers through the thickness of the PZT matrix. Magnetic-force microscopy images of the nanocomposite thin film under opposite magnetic poling conditions revealed in-plane pancake-like regions of higher concentration of the CFO nanoparticles. The constraints on the size and distribution of the CFO nanoparticles created a unique distribution in a PZT matrix and achieved values of ME coupling of 3.07 V cm-1 Oe-1 at a DC bias of 250 Oe and 1 kHz, increasing up to 25.0 V cm-1 Oe-1 at 90 kHz. Piezo-force microscopy was used to investigate the ferroelectric domain structure before and after opposite magnetic poling directions. It was found that in this nanocomposite, the polarization of the ferroelectric domains switched direction as a result of switching the direction of the magnetization by magnetic fields.