ML-based interatomic potentials for accelerated materials simulations, including neural network potentials, Gaussian approximation potentials, and equivariant models.

GANs, diffusion models, VAEs, and transformer-based architectures for inverse design, crystal structure prediction, and de novo molecule generation.

DFT-based and ML-accelerated virtual screening of large chemical spaces, materials databases, and automated workflow pipelines.

Self-driving labs, automated synthesis platforms, robotic experimentation, and closed-loop AI-driven optimization of materials properties.

Integrating text, images, spectral data, and simulation results with large language models, vision-language models, and multimodal foundation models for materials science.

Embedding physical constraints, symmetries, and conservation laws into ML models; Bayesian methods and conformal prediction for reliable materials property prediction.

Applications to batteries, catalysts, photovoltaics, thermoelectrics, CO2 capture, and hydrogen storage materials.