Data-driven, low-dimensional, and generative models for molecular and materials discovery and design

University of Chicago Center in Paris, 39 Rue des Grands Moulins, Paris, France
June 16-18, 2025
Organizers: Andrew Ferguson (UChicago), Jérôme Hénin (IBPC, CNRS)

Workshop Scope

A persistent challenge in molecular modeling is the prevalence of so-called “rare event-dynamics”, wherein the system becomes trapped for long periods of time in particular states from which it only infrequently escapes. The fundamental origin of this challenge is a mismatch in the slow time scales for these phenomena (seconds, minutes, hours) relative to the longest time scales accessible to simulations even on the world’s fastest supercomputers (μs-ms). This discrepancy presents a challenge for proper sampling of the system states and, therefore, the accurate prediction of its properties. For example, a simulation of an enzyme may become trapped in a misfolded state and never transition to the active state that catalyzes a biochemical reaction, or vacancies in a crystal lattice responsible for charge transport in solar cells may never hop to new lattice sites, compromising the accuracy of the molecular simulation predictions of photovoltaic performance.

Enhanced sampling techniques present a means to break this so-called “time-scale barrier” by offering a variety of ways to synthetically accelerate the system dynamics. These approaches typically apply artificial biasing forces, modify the underlying interaction potentials, or focus sampling on a particular transition of interest, and then apply analytical or numerical corrections for the influence of the known biases. In recent years, artificial intelligence (AI) and machine learning (ML) methods have opened new capabilities in enhanced sampling by providing unprecedented levels of power and control in learning collective variables, applying flexible biasing forces, or even enabling completely new methodologies. In general, the most powerful and exciting techniques have been adaptations of tools developed by ML practitioners for general purpose applications such as object recognition, automated captioning, image inpainting, or natural language processing. In one sense, the adaptation of techniques is an exceedingly powerful paradigm in transferring new tools and concepts to invigorate the field, but, in another, the absence of physical laws, thermodynamic constraints, and domain knowledge has limited the capabilities and effectiveness of these methodologies and the interpretability of the results.

The effective integration of new AI and ML tools into the field of molecular simulation requires an interdisciplinary synthesis of state-of-the-art foundational and generic AI/ML tools (e.g., denoising diffusion probability models, flow matching, Bayesian optimization) with fundamental molecular modeling domain knowledge and constraints (e.g., energy conservation, force equivariance, molecular synthesizability). It is the primary motivation for this workshop to convene leading researchers in applied machine learning, theoretical chemistry, statistical thermodynamics, and biophysics to discuss and share new discoveries and tools, identify open challenges and pressing needs in the field, and establish new disciplinary bridges and collaborative opportunities. Specific technical topics of interest include data-driven and low-dimensional techniques, generative models, Physics-Aware AI (PAI), and eXplainable AI (XAI), along both foundational and applied dimensions. An additional primary focus of the workshop will be consideration of sustainable modeling and simulation techniques in how novel AI/ML tools can help alleviate the large electrical and cooling costs associated with large-scale molecular simulations on supercomputing and cloud hardware, and mitigate the negative impact that these calculations may themselves have on the climactic and environmental challenges that they seek to help solve through the discovery of new molecules and materials.

Location and Venue

The workshop will be held at the University of Chicago Paris Center, located 39 Rue des Grands Moulins in the 13th arrondissement of Paris. The Center is designed by acclaimed architects, Studio Gang, and features several venues for conferences, symposia, and events, a research suite, classrooms of varying size and purpose, a small library, and recreational spaces for the Center’s faculty, students, staff, and visitors.

Financial Support

This workshop is co-sponsored by the University of Chicago and CECAM-US-CENTRAL, the first CECAM node in the United States, and will be held at the UChicago International Institute of Research in Paris (IIRP) in Paris, France. There are no registration fees for this event, and we are pleased to be able to offer invited speakers from outside of Paris limited financial support to defray the costs of travel and accommodation. For invited speakers traveling from within Europe (but outside of Paris), we can provide a total of $600 per person. For invited speakers traveling from outside of Europe, we can provide a total of $900 per person.

Travel and Accommodation

Paris is well served by international and domestic flights through Charles de Gaulle (CDG) international airport. The Paris Center is easily accessible from the airport by public transport taking approximately 50 minutes on RER or Metro. Paris is also extremely well served by rail service via six major train stations: Gare du Nord, Gare de Lyon, Gare de l’Est, Gare Montparnasse, Gare d’Austerlitz, and Gare Saint-Lazare.

A number of hotels are conveniently located in close proximity to the UChicago Paris Center. June is high-season for hotels in Paris so rates may be elevated. In some cases, special rates are available to participants.

1) Ibis Hotel Bercy Village, 19 Pl. des Vins de France, 75012 Paris (10% off published rates by mentioning “University of Chicago”)
2) Pullman Paris Centre Bercy, 1 Rue de Libourne, 75012 Paris (special rate of 265€ per night by mentioning “University of Chicago”)
3) Too Hotel, 65 Rue Bruneseau, 75013 Paris (brand new modern hotel designed by Philip Stark, special rate of 300€ per night by mentioning “University of Chicago”)
4) Green Hotels Paris 13, 90 Rue de Patay, 75013 Paris (within walking distance)
5) Albe Hotel Bastille, 66 Rue de Charenton, 75012 Paris

Local Restaurants

A number of excellent and inexpensive sit-down and to-go options are available in close proximity to the Center.

Sit-down lunch and dinner options:

1) Nosso, 22 Prom. Claude Lévi-Strauss, 75013 Paris (For lunch, between 50€ and 80€ per person. For dinner, between 65€ and 120€ per person. Does not cater to groups.)
2) Tempero Épicerie, 24 Prom. Claude Lévi-Strauss, 75013 Paris (For lunch, between 30€ and 50€ per person.)
3) Anco, 108 Rue de Bercy, 75012 Paris (For lunch, between 35€ and 50€ per person. For dinner, between 55€ and 80€ per person.)
4) L’Auberge Aveyronnaise, 40 Rue Gabriel Lamé, 75012 Paris (For lunch, between 35€ and 50€ per person. For dinner, between 55€ and 80€ per person.)
5) Milord, 12 Rue de Tolbiac, 75013 Paris (For lunch, between 35€ and 50€ per person. For dinner, between 40€ and 70€ per person.)
6) Dupont Café Bibliothèque, 84 Av. de France, 75013 Paris (For lunch, between 20€ and 30€ per person. Right across from Paris Center.)

To-go lunch options:

1) Kayser boulangerie, 77 Quai Panhard et Levassor, 75013 Paris (Lunch box is around 12€.)
2) Prêt-à-manger, 110 Av. de France, 75013 Paris (Lunch box is around 12€.)
3) Exki, 116 Av. de France, 75013 Paris (Lunch box is around 12€.)

Organizers

Jérôme Hénin IBPC, CNRS jerome.henin@cnrs.fr
Andrew Ferguson UChicago andrewferguson@uchicago.edu 

Administrative Support

Maria Guarte UChicago Paris Center mmguarte@uchicago.edu 

Invited Speakers

Aaron Dinner UChicago dinner@uchicago.edu 
Andrew Ferguson UChicago andrewferguson@uchicago.edu 
Benoit Roux UChicago roux@uchicago.edu 
Bernd Ensing U Amsterdam B.Ensing@uva.nl 
Bettina Keller FU Berlin bettina.keller@fu-berlin.de 
Christine Peter U Konstanz christine.peter@uni-konstanz.de 
Christophe Chipot U Lorraine Christophe.Chipot@Univ-Lorraine.fr 
Clément Wespiser CEA-DAM    Clement.WESPISER@cea.fr 
Fabio Pietrucci Sorbonne fabio.pietrucci@sorbonne-universite.fr 
Gabriel Stoltz Ecole des Ponts gabriel.stoltz@enpc.fr
Jérôme Hénin IBPC, CNRS jerome.henin@cnrs.fr 
Juan de Pablo UChicago depablo@uchicago.edu 
Jean-Philip Piquemal Sorbonne jean-philip.piquemal@sorbonne-universite.fr 
Julien Lam Université de Lille julien.lam@cnrs.fr 
Marylou Gabrié Ecole Polytechnique marylou.gabrie@polytechnique.edu 
Paraskevi Gkeka Sanofi paraskevi.gkeka@sanofi.com 
Suri Vaikuntanathan UChicago svaikunt@uchicago.edu 
Tony Lelièvre Ecole des Ponts tony.lelievre@enpc.fr

Schedule

Day 1 –
Monday June 16, 2025
8:30-8:50REGISTRATION AND COFFEELounge
8:50-9:00WELCOME & INTRODUCTIONS
Andrew Ferguson (UChicago)
Jérôme Hénin (CNRS)
Hugo Toudic & Maria Guarte (IIRP)
Amphitheater
9:00-9:40Jérôme Hénin
“TBA”
Amphitheater
9:40-10:20Bernd Ensing
“TBA”
Amphitheater
10:20-10:50COFFEE BREAKLounge
10:50-11:30Suri Vaikuntanathan
“Active generative diffusion”
Amphitheater
11:30-12:10Gabriel Stoltz
“A mathematical analysis of autoencoders for (un)supervised dimensionality reduction”
Amphitheater
12:10-14:00LUNCH (ON YOUR OWN)Local Restaurant
14:00-14:40Benoit Roux
“TBA”
Amphitheater
14:40-15:20Paraskevi Gkeka
“Targeting RNA with small molecules”
Amphitheater
15:30-16:00LIGHTNING TALKSAmphitheater
16:00-17:30POSTER SESSIONGreat Room
17:30-19:00COCKTAIL HOURGreat Room
Day 2 –
Tuesday June 17, 2025
8:30-9:00REGISTRATION AND COFFEELounge
9:00-9:40Andrew Ferguson
“TBA”
Amphitheater
9:40-10:20Clément Wespiser
“Data-driven design of synthesizable high energy materials”
Amphitheater
10:20-10:50COFFEE BREAKLounge
10:50-11:30Julien Lam
“Harvesting nucleating structures in first order transitions”
Amphitheater
11:30-12:10Marylou Gabrié
“TBA”
Amphitheater
12:10-14:00LUNCH (ON YOUR OWN)Local Restaurant
14:00-14:40Juan de Pablo
“Data-Driven Design of Polymeric Molecules and Materials”
Amphitheater
14:40-15:20Jean-Philip Piquemal
“TBA”
Amphitheater
15:30-17:00POSTER SESSIONGreat Room
19:00-21:00SPEAKER DINNERTBA
Day 3 –
Wednesday June 18, 2025
8:30-9:00REGISTRATION AND COFFEELounge
9:00-9:40Tony Lelièvre
“TBA”
Amphitheater
9:40-10:20Bettina Keller
“Thermal isomerization of retinal: reaction coordinate and committor function”
Amphitheater
10:20-10:50COFFEE BREAKLounge
10:50-11:30Aaron Dinner
“Graph neural networks for analyzing dynamics”
Amphitheater
11:30-12:10Christine Peter
“Biomolecular residue interaction landscapes analyzed via temporal network embedding”
Amphitheater
12:10-14:00LUNCH (ON YOUR OWN)Local Restaurant
14:00-14:40Christophe Chipot
“TBA”
Amphitheater
14:40-15:20Fabio Pietrucci
“TBA”
Amphitheater
15:20-15:30CLOSING REMARKS
Andrew Ferguson (UChicago)
Jérôme Hénin (CNRS)
Amphitheater