Machine-learning approach for discovery of conventional superconductors

First-principles computations are the driving force behind numerous discoveries of hydride-based superconductors, mostly at high pressures, during the last decade. Machine-learning (ML) approaches can further accelerate the future discoveries if their reliability can be improved. The main challenge of current ML approaches, typically aiming at predicting the critical temperature $T_{\rm c}$ of a solid from its chemical composition and target pressure, is that the correlations to be learned are deeply hidden, indirect, and uncertain. In this work, we showed that predicting superconductivity at any pressure from the atomic structure is sustainable and reliable. For a demonstration, we curated a diverse dataset of 584 atomic structures for which $\lambda$ and $\omega_{\log}$, two parameters of the electron-phonon interactions, were computed. We then trained some ML models to predict $\lambda$ and $\omega_{\log}$, from which $T_{\rm c}$ can be computed in a post-processing manner. The models were validated and used to identify two possible superconductors whose $T_{\rm c}\simeq 10-15$K at zero pressure. Interestingly, these materials have been synthesized and studied in some other contexts. In summary, the proposed ML approach enables a pathway to directly transfer what can be learned from the high-pressure atomic-level details that correlate with high-$T_{\rm c}$ superconductivity to zero pressure. Going forward, this strategy will be improved to better contribute to the discoveries of new superconductors.