Quantum Transitions

Machine learning is a powerful tool to analyze complex data, but can it help reveal unexplored domains of knowledge? We answer this question in the affirmative, showing in this work that one can predict phase transitions using Gaussian process extrapolation across parameter space.

I am interested in understanding the manifestations of quantum interactions as transitions in physical observables in experiment. This includes quantum phase transitions and polaron transitions. The former involves the cooperative behavior of a large collection of particles driven by quantum correlations. The latter refers to situations in which a polaronic quasiparticle exhibits a transition as a function of the coupling strength at which point two energy levels cross, or, more interestingly, the ground state changes character abruptly. Besides their importance to foundational theory, these transitions often serve as an attractor to other emergent behavior. For example, superconductivity emergent near a ferroelectric-paraelectric transition in STO represents one class of such problems.