Pavan Subramani wins $25,000 Davidson scholarship for AI drug discovery
Pavan Subramani, a 17-year-old from West Virginia, just scored a $25,000 Davidson scholarship for his AI project that could seriously speed up drug discovery.
By mixing generative diffusion models with molecular dynamics simulations, his approach makes studying protein movements and drug interactions in tests involving alanine dipeptide for dhMD 68 times more efficient in terms of the effective sample size per GPU hour for the slow phi dihedral.
dhMD and dhBLUES accelerate drug studies
Pavan's new technique, called diffusion hybridized molecular dynamics (dhMD), helps computers predict protein changes much faster without losing accuracy.
He also created a version for testing how drugs bind to proteins (dhBLUES), which converged to equilibrium ligand-binding populations more than four times faster than BLUES.
His work could make advanced drug research faster and more accessible to academic laboratories with limited computing resources.