using scanning tunneling spectroscopy and machine learning*

Pyrochlore iridates have attracted great interest due to their geometrically frustrated lattice with potential for realizing metallic spin liquid behavior. Pr₂Ir₂O₇ is uniquely located near the quantum critical point of antiferromagnetic metal to insulator transition. We carried out STM experiments on the [111] Kagome surface of the pyrochlore iridate Pr₂Ir₂O₇. STM topographs show the six-fold atomic structure expected from the [111] surface demonstrating atomic resolution STM images on a pyrochlore iridate. At low temperature, the tunneling density of states reveal Kondo behavior with Fano lineshape near the Fermi energy that exhibit strong electronic inhomogeneity. The large quantity of generated data and the intrinsic electronic inhomogeneity in this system introduce enough complexity to render conventional data analysis inadequate. Using machine learning of large data, we discover these rather seemingly random spectra to form nanoscale patterns, which may be related to electronic entanglement.