Modeling and Simulation Group Meeting
A Festival of Machine Learning in Science at NYU (Part 1)
Speaker: David Grier, Glen Hocky, Mark Tuckerman
Location: TBA
Date: Thursday, April 1, 2021, 12:30 p.m.
Synopsis:
A series of short talks on recent applications of machine learning by a collection of speakers from chemistry, physics, and math.
Part I:
1. David Grier (Physics)
"CATCH: Characterizing and Tracking Colloids Holographically using Convolutional Neural Networks"
CATCH is a system of convolutional neural networks that analyzes experimentally recorded
holograms of colloidal particles to measure the particles' three-dimensional positions,
their diameters and their refractive indexes. The CNN implementation is faster and more robust
than conventional hologram-analysis algorithms and has real-world applications in
pharmaceutical manufacturing, medical testing and semiconductor processing.
2. Glen Hocky (Chemistry)
"Coarse-grained directed simulation"
This talk will present an on-the-fly maximum likelihood approach to matching experimental or other observables within a molecular dynamics simulation.
3. Mark Tuckerman (Chemistry)
"Driving solid-solid phase transitions in crystalline materials with machine learning"
Predicting the kinetics of transitions between different solid phases of crystalline materials
remains a significant challenge. Here, we introduce an approach that employs neural networks
to drive enhanced sampling simulations in a classification space that is capable of reproducing
mechanisms and kinetic barriers associated with structural transition events.