Fig. 1: General workflow of our work. | Nature Communications

Fig. 1: General workflow of our work.

From: Machine learning-assisted dual-atom sites design with interpretable descriptors unifying electrocatalytic reactions

Fig. 1

It includes four major parts: (i) primitive descriptor (ϕxx) for atomic property effects through d-band shape analysis, (ii) screening principle (ϕopt) of potential desirable heteronuclear DACs based on reactant effects, (iii) ML-based descriptors (ϕxy) for synergistic effects through physically meaningful feature engineering based on ϕxx and feature selection/sparsification algorithms, (iv) the final universal descriptor model (Φ) with quantification of coordination effects and corresponding experimental verifications.

Back to article page