Track synthesis, materials and catalysis as it is published.
Search papers, labs, and topics across Lattice.
Affiliation:
2
0
0
11
Algorithmic differentiation eliminates the nightmare of deriving analytical second derivatives for meta-GGAs, unlocking fully differentiable perturbation theory for end-to-end training of neural-network exchange-correlation functionals.
Standard uncertainty quantification fails downstream in atomistic ML because it destroys physical error cancellations, a blind spot resolved by multivariate conformal methods that explicitly preserve cross-property correlations.