Trustability and physical interpretability are critical requirements for the development of robust and sustainable machine learning systems needed for accelerated science. This research opportunity focuses on developing, evaluating, and applying computational methods for materials characterization and/or simulation that combine the best aspects of physics based and machine learning based modeling. The advanced materials characterization techniques focus on automated phase identification and quantitative analysis from laboratory and synchrotron x-ray diffraction, x-ray absorption spectroscopy for the quantification of chemical short range order, and automated microstructural image analysis. The simulation approaches of interest include machine learning force fields and probabilistic multiscale/multiphysics surrogate modeling. Preferred areas of application include structure and microstructure of metallic alloys and relationships between defects and properties in semiconductor device materials. This research is expected to be executed through internal and external collaborations with comprehensive expertise in advanced material simulation, high throughput experimentation, and automated laboratory systems.