Special Seminars
Speaker: Fang Liu, Assoc. Prof., Emory University
Date: Monday, September 21, 2026 at 2-3 pm
Location: Noyes Lab 147
Faculty host: Garnet Chan
Title: Synergizing Quantum Chemistry and Machine Learning for Molecular Discoveries in the Condensed Phase
Abstract:
Machine learning (ML) and big data play increasingly critical roles in chemical discovery. However, datasets and ML models for condensed-phase molecular systems, such as solvated molecules and molecule assemblies, remain scarce. My research group leverages GPU-accelerated quantum chemistry and machine learning to address these gaps.
Many crucial solvent-solute interactions cannot be captured by the implicit solvent models routinely used in quantum chemistry calculation and require explicit solvent treatment. To streamline the simulation workflow for explicit-solvent calculations, we developed AutoSolvate, an open-source toolkit, which was later developed into a chatbot-assisted, cloud-based platform that automates simulation setup and execution using cloud computing resources. These tools have enabled the efficient generation of computational datasets for solvated molecules, which were used to train Δ-ML models to enhance the accuracy of low-cost computational methods against experimental measurements. For molecular assemblies, we developed a size-transferable machine-learned exciton model to rapidly model their excited-state properties. More recently, we developed a domain-informed multi-agent AI framework that autonomously generates quantum-chemical training data, constructs molecule-specific machine-learned Hamiltonians and potential energy surfaces, and performs excited-state dynamics, with independent agents verifying scientific validity throughout the workflow. Additionally, we aim to bridge the gap between simulated and experimental datasets by leveraging large volumes of computational data to train ML models for real-time analysis in autonomous experiments. As a proof of concept, we successfully trained an ML model to detect material phase transitions in situ using angle-resolved photoemission spectroscopy (ARPES).
These events are open to the Caltech/JPL community members only. No recordings are allowed except by permission of the organizer.
Speaker: Joshua S. Kretchmer, Asst. Prof., Georgia Institute of Technology
Date: Tuesday, April 28, 2026 at 4-5 pm
Location: Noyes Lab 147
Faculty host: Garnet Chan
Title: Real-time simulations of ultrafast electron relaxation dynamics
Abstract:
The Kretchmer Group develops simulation methods at the intersection of electronic structure theory and quantum dynamics. We apply our methods to examine electron and spin dynamics, non-adiabatic processes, and charge transport in complex molecular and material systems including weakly-bound molecular clusters, chiral and helical molecules, and perovskites. In this talk, I will highlight our recent development of a new methodology to unravel the competing electronic relaxation pathways following inner-shell ionization in weakly bound systems. When an inner-shell electron is removed, another electron relaxes to fill the vacancy, which can transfer energy to a nearby entity, generating a low-energy secondary ionized electron. Critically, the location of the electron that relaxes and the secondary ionized electron can originate from either the original molecule or a neighboring molecule. The relaxation can also compete with electron, hole, and even proton transfer. Our work combines real-time DFT with a complex absorbing potential to simulate the explicit dynamics of this complex array of processes. We have additionally introduced coupled electron-nuclear dynamics through Ehrenfest dynamics, which enabled an in-depth investigation of the fragmentation pathways in the water dimer following inner-valence ionization. Our simulations identified a previously unknown relaxation channel essential to reproducing the experimentally observed fragmentation products of our collaborators, prompting a reinterpretation of the origin of low-energy electrons in hydrogen-bonded systems such as water.
These events are open to the Caltech/JPL community members only. No recordings are allowed except by permission of the organizer.