6 papers
BERTology of Molecular Property Prediction
Mohammad Mostafanejad, Paul Saxe, T. Daniel Crawford
Chemical language models (CLMs) have emerged as promising competitors to popular classical machine learning models for molecular property prediction (MPP) tasks. However, an increa…
SEAMM: A Simulation Environment for Atomistic and Molecular Modeling
Paul Saxe, Jessica Nash, Mohammad Mostafanejad +4
The Simulation Environment for Atomistic and Molecular Modeling (SEAMM) is an open-source software package written in Python that provides a graphical interface for setting up, exe…
Numerical Analysis of the Complete Active-Space Extended Koopmans's Theorem
Reza Hemmati, Mohammad Mostafanejad, J. V. Ortiz
We investigate the numerical accuracy of the extended Koopmans's theorem (EKT) in reproducing the full configuration interaction (FCI) and complete active-space configuration inter…
Unification of popular artificial neural network activation functions
Mohammad Mostafanejad
We present a unified representation of the most popular neural network activation functions. Adopting Mittag-Leffler functions of fractional calculus, we propose a flexible and com…
Global hybrid multiconfiguration pair-density functional theory
Mohammad Mostafanejad, Marcus Dante Liebenthal, A. Eugene DePrince
A global hybrid extension of variational two-electron reduced-density matrix (v2RDM)-driven multiconfiguration pair-density functional theory (MCPDFT) is developed. Using a linear…
Combining pair-density functional theory and variational two-electron reduced-density matrix methods
Mohammad Mostafanejad, A. Eugene DePrince
Complete active space self-consistent field (CASSCF) computations can be realized at polynomial cost via the variational optimization of the active-space two-electron reduced-densi…