
David Williams
28
Documents
Identifiants chercheurs
Présentation
ORCID: 0000-0001-9627-5647 Web of Science: AFR-4512-2022
Past Employment
Postdoctoral researcher with a focus on developing efficient artificial neural network based approaches for accurate, high dimensional excited state dynamics.
- Sep 2023 – Present European MCSA-Bienvenue postdoctoral fellow at the Institute of Physics Rennes (IPR), France
- Spring 2023 collaborating visit at the IPR, Rennes, France
- Year 2022 postdoctoral researcher in the Martínez group at Stanford University, CA, USA
- Year 2021 postdoctoral reseacher in the Eisfeld group at the University of Bielefeld, Germany
- Jan 2021 PhD under Wolfgang Eisfeld at the University of Bielefeld, Germany
Research Interests
- Photochemical processes in atmospheric and astrochemistry: Particular focus on isolated molecules in the gas phase at finite temperature, with research experience in the modelling of photodetachment processes.
- Numerical method development based on artificial neural networks: Developed and refined schemes for the construction of diabatic potential energy surfaces using artificial neural networks. Particularly focused on preconditioning techniques to imbue models with physical information, reducing the strain of the network.
- Efficient, modularized models of complex systems: Continued efforts to provide schemes that allow for models to be extended to the full-dimensional problem by building upon restricted-dimensional models to mitigate innate problems of high-dimensional coordinate spaces (curse of dimensionality).
- Molecular quantum dynamics of excited states: Conducted in-depth theoretical studies of Jahn-Teller type systems and the impact of the geometric phase effect on their spectroscopy.
- Mechanochemistry and photoswitches: General interest in the development of high quantum yield photoswitches and molecular machines.
Education
- 2016 – Jan. 2021 Dr. rer. nat. (PhD), Theoretical Chemistry, Bielefeld University, Germany
- "Conquering coupled diabatic potential energy surfaces with artificial neural networks for nonadiabatic dynamics''
- 2014 – 2016 Master of Science, Theoretical Chemistry, Bielefeld University, Germany
- "Neural Network Based Methods for the Representation of Diabatic Potential Energy Surfaces''
Compétences
Excited state modeling
Nonadiabatically coupled potential method development
Artificial neural network based models
Publications
6
2
1
1
9
3
4
2
2
2
3
1
1
Publications
A "post vibronic coupling ansatz" for accurate diabatic potentialsTHEMOSIA Western Pole scientific day (2024), Dec 2024, Rennes, France
Communication dans un congrès
hal-04822824
v1
|
|
Accurate coupled potential energy surfaces: transcending vibronic coupling with artificial neural networksInvited seminar, CIMAP, Nov 2024, Caen, France
Communication dans un congrès
hal-04830615
v1
|
|
Beyond Ultra-Fast Processes: Accurate Coupled Potential Energy SurfacesInvited seminar, Strasbourg Institute of Chemistry, Nov 2024, Strasbourg, France
Communication dans un congrès
hal-04830634
v1
|
|
Analytical and Numerical Pathways Towards Understanding Nonadiabatic DynamicsInvited seminar, Institut Charles Gerhardt Montpellier, Dec 2023, Montpellier, France
Communication dans un congrès
hal-04718452
v1
|
|
Efficient Nonadiabatic Dynamics with Neural Network Based Coupled Potential Energy SurfacesInvited Seminar, CEISAM Lab, Mar 2023, Nantes, France
Communication dans un congrès
hal-04718438
v1
|
|
Artificial Neural Networks: Toward Efficient Nonadiabatic Dynamics with Coupled Potential Energy SurfacesAmerican Conference on Theoretical Chemistry, https://www.actc2022.com/, Jul 2022, Palisades Tahoe, United States
Communication dans un congrès
hal-03976934
v1
|
|
Dynamics on accurate coupled artificial neural network based potential surfaces : NO_3^- photodetachment spectra7th High Dimensional Quantum Dynamics (HDQD), https://www.hdqd2022.org/, Jul 2022, Groningen, Netherlands
Communication dans un congrès
hal-02262273
v1
|
|
Photodetachment spectrum of NO3-: Jahn-Teller effect and artificial neuron networkCECAM2021 : Modeling dilute matter under ultracold solvent: from quantum structure to dynamics and reactivity, May 2021, Grenoble, France
Communication dans un congrès
hal-03212344
v1
|
|
Symmetry, geometric phase and spectroscopyElectronic Structure, Dynamics & Spectroscopy (Domcke's 70th Birthday), Apr 2018, Garching, Germany
Communication dans un congrès
hal-01773875
v1
|
The Molecular Foucault Pendulum: What is a "Geometric Phase"?ERRIE “European Responsible Research and Innovation Event” 2024, Oct 2024, Rennes, France
Poster de conférence
hal-04830735
v1
|
|
Neural Network coupled potential and dipole surfaces for interstellar molecules24th European Conference on the Dynamics of Molecular Systems - MOLEC2024, Jun 2024, Aarhus, Denmark
Poster de conférence
hal-04597242
v1
|
|
Representing non-A_1 Operators:C_nv Diabatic Dipole Transition SurfacesTHEMOSIA Western Pole scientific day (2024), Dec 2024, Rennes, France
Poster de conférence
hal-04816579
v1
|
|
Diabatic potential energy surfaces for Renner-Teller systems using Artificial Neural Networks8th High Dimensional Quantum Dynamics (HDQD), Jul 2024, Hamburg, Germany
Poster de conférence
hal-04579831
v1
|
|
On the vibronic level structure of cubic point groups8th High Dimensional Quantum Dynamics (HDQD), Jul 2024, Hamburg, Germany
Poster de conférence
hal-04579829
v1
|
|
Representing non-A_1 Operators:C_nv Diabatic Dipole Transition Surfaces8th High Dimensional Quantum Dynamics (HDQD), Jul 2024, Hamburg, Germany
Poster de conférence
hal-04579830
v1
|
|
Efficient Nonadiabatic Dynamics with Artificial Neural NetworksAmerican Conference on Theoretical Chemistry, Jul 2022, Palisades Tahoe, United States
Poster de conférence
hal-03976949
v1
|
|
Accurate vibronic quantum dynamics of NO3 using a diabatic neural network potential22nd European Conference on the Dynamics of Molecular Systems, Aug 2018, Dinard, France
Poster de conférence
hal-03976972
v1
|
|
Neural Network Diabatization: A new ansatz for high-dimensional coupled potential energy surfacesSymposium on Theoretical Chemistry, Aug 2017, Basel, Switzerland
Poster de conférence
hal-03976989
v1
|
|
Conquering coupled diabatic potential energy surfaces with artificial neural networks for nonadiabatic dynamicsChemical Sciences. University of Bielefed, 2020. English. ⟨NNT : ⟩
Thèse
tel-03917774
v1
|
Chargement...
Chargement...