Abstract In this paper, we demonstrate an application of the Transformer self-attention mechanism in context materials science. Our network, Compositionally Restricted Attention-Based network (), explores area structure-agnostic property predictions when only a chemical formula is provided. results show that ’s performance matches or exceeds current best-practice methods on nearly all 28 total ...

Two-phase random textures abound in a host of contexts, including porous and composite media, ecological structures, biological media, and astrophysical structures. Questions surrounding the spatial structure of such textures continue to pose many theoretical challenges. For example, can two-point correlation functions be identified that can be manageably measured and yet reflect nontrivial hig...

Skin-based wearable devices have a great potential that could result in a revolutionary approach to health monitoring and diagnosing disease. With continued innovation and intensive attention to the materials and fabrication technologies, development of these healthcare devices is progressively encouraged. This article gives a concise, although admittedly non-exhaustive, didactic review of some...

Materials science offers microbiologists a wide variety of organic and inorganic materials with chemical and physical properties that can be precisely controlled. These materials present new capabilities for isolating, manipulating and studying bacteria and other microorganisms and are poised to transform microbiology. This review summarizes three classes of materials that span a range of lengt...

For cost effective machining it is necessary to identify and quantify changes in tool material machining characteristics. Cryoprocessing, a supplementary process to conventional heat treatment, involves deepfreezing of materials at cryogenic temperatures to enhance the mechanical and physical properties. The execution of cryoprocessing on cutting tool materials increases wear resistance, hardne...