Machine learning-enabled prediction of chemical durability of A2B2O7 pyrochlore and fluorite

•A database of long-term and short-term elemental leaching rates of A2B2O7-structures and their correlations with key materials parameters and structural characteristics are established.•A machine learning approach has been applied to elucidate the key parameters governing the leaching behaviors of A2B2O7-structures.•Machine learning models based linear regression and KRR are developed for a data-driven prediction of materials’ chemical durability with high fidelity.•Demonstration of the machine-learning approach to accelerate design and discovery of materials with optimized compositions and performance for nuclear waste form applications.