AFLOW-ML: A RESTful API for machine-learning predictions of materials properties

摘要: Machine learning approaches, enabled by the emergence of comprehensive databases materials properties, are becoming a fruitful direction for analysis. As result, plethora models have been constructed and trained on existing data to predict properties new systems. These powerful methods allow researchers target studies only at interesting $\unicode{x2014}$ neglecting non-synthesizable systems those without desired thus reducing amount resources spent expensive computations and/or time-consuming experimental synthesis. However, using these predictive is not always straightforward. Often, they require panoply technical expertise, creating barriers general users. AFLOW-ML (AFLOW $\underline{\mathrm{M}}$achine $\underline{\mathrm{L}}$earning) overcomes problem streamlining use machine developed within AFLOW consortium. The framework provides an open RESTful API directly access continuously updated algorithms, which can be transparently integrated into any workflow retrieve predictions electronic, thermal mechanical properties. types interconnected cloud-based applications envisioned capable further accelerating adoption development.