Prediction of vibrational spectra by machine Learning and algorithmic graph theory

We are seeking for a candidate who will apply for a 48 months PhD funded by the China Scholarship Council (CSC) at the Univ Paris-Saclay. The candidate has to be of Chinese nationality and has to be currently enrolled in a Master M2 degree in a Chinese university or at the Univ Paris-Saclay, with her/his diploma to be obtained any time during the year 2022. Please see the details for the application and recruitment requirements on the Paris-Saclay website: https://www.universite-paris-saclay.fr/csc-programme-doctoral-conjoint-de-luniversite-paris-saclay-et-du-china-scholarship-council

We are seeking for a candidate with a background in either theoretical informatics with expertise in machine learning methods and/or algorithmic graph theory, or in computational theoretical chemistry, or in computational theoretical physics. For chemistry and physics backgrounds, knowledge in machine learning techniques is a plus, as well as knowledge in molecular dynamics MD simulations. Depending on the background/expertise of the candidate, she/he will become expert in the complementary topic thanks to the bi-disciplinary team of PhD supervisors. Skills in coding (e.g. C, fortran, python, others) is a plus.

We are a bi-disciplinary team with expertise in molecular dynamics (MD) simulations for vibrational spectroscopy and in algorithmic graph theory and machine learning approaches. Our group is internationally recognized for theoretical anharmonic vibrational spectroscopy calculations via finite temperature DFT-based molecular dynamics simulations. See some of our reviews and recent high impact papers on the topic,^1-7 with applications to IR/Raman/SFG spectroscopies for all states of matter (gas phase molecules & clusters, liquids, solid-liquid interfaces). We are also recognized experts in algorithmic graph theory and machine learning approaches. See two of our recent works on graph theory.^8-9

We are now developing new theoretical tools that will predict both the spectrum and the associated molecular 3D-structure responsible for the spectroscopic fingerprints, without the aid of computational chemistry calculations. The methodology is developed based on machine learning and deep learning approaches, associated to algorithmic graph theory and game theory.

This is the goal of the PhD funded by CSC.

For discussions and more details on the PhD project, please contact Prof Marie-Pierre Gaigeot at the University Paris-Saclay, Campus of the University of Evry: [email protected]

[1] Opinion paper on MD-based vibrational spectroscopy of gas phase molecules and their assembly, Spectrochimica Acta A 260:119864-883 (2021)

[2] Review paper on DFT-MD for anharmonic spectroscopy, Phys. Chem. Chem. Phys. 12:3336-59 (2010)

[3] Molecular fingerprints of hydrophobicity at aqueous interfaces from theory and vibrational spectroscopies, J. Phys. Chem. Letters 12:3827-36 (2021)

[4] Stripping away ion hydration shells in electrical double layer formation: water networks matter

PNAS 118:e2108568118 (2021)

[5] Ions Tune Interfacial Water Structure and Modulate Hydrophobic Interactions at Silica Surfaces.
J. Am. Chem. Soc. 142:6991-7000 (2020)

[6] Deconvolution of BIL-SFG and DL-SFG Spectroscopic Signals Reveal Order/Disorder of Water at the Elusive Aqueous Silica Interface. Phys. Chem. Chem. Phys. 21:22188-22202 (2019)

[7] Molecular Hydrophobicity at a Hydrophilic Surface. PNAS 116:1520-1525 (2019)

[8] Graph theory for automatic structural recognition in molecular dynamics simulations. J. Chem. Phys. 149:184102-15 (2018)

[9] Conformational assignment of gas phase peptides and their H-Bonded complexes using far-IR/THz: IR-UV ion dip experiment, DFT-MD spectroscopy, and Graph Theory for modes assignment. Faraday Discussions, 217:67 (2019)