tbc
\r\n
\r\nAgenda
\r\n12:10 – 12:30 Session
\r\n12:30 – 14:00 Standing Lunch
The ISIC Faculty retreat will be replaced by an on-site ISIC Council followed by a lunch and Faculty discussion.
\r\n
\r\nDraft Agenda
\r\n09:00 – 9:30 Welcome Coffee
\r\n09:30 – 11:00 ISIC Council
\r\n11:00 – 12:00 Faculty Discussion
\r\n12:00 – 13:30 Standing lunch
\r\n13:30 – 17:00 Faculty Discussion
\r\n
\r\n
You can apply to participate and find all the relevant information (speakers, abstracts, program,...) on the event website: https://www.cecam.org/workshop-details/from-data-to-dynamics-machine-learning-in-statistical-mechanics-and-molecular-simulations-1487.
\r\n
\r\nRegistration is required to attend the full event, take part in the social activities and present a poster at the poster session (if any). However, the EPFL community is welcome to attend specific lectures without registration if the topic is of interest to their research. Do not hesitate to contact the CECAM Event Manager if you have any question.
\r\n
\r\nDescription
\r\nSince its introduction in the 1970s, molecular dynamics (MD) has become an indispensable computational microscope for studying complex biological systems at atomic resolution. It has enabled detailed investigations into protein folding, conformational dynamics, and ligand binding and unbinding. Over the past decade, increasing computational power has made microsecond-scale simulations routine, producing massive datasets that demand sophisticated analysis strategies [1]. Despite these advances, conventional MD simulations still face a fundamental limitation: many biologically relevant events occur over milliseconds to seconds—timescales largely inaccessible to standard MD.
\r\nTo bridge this gap, researchers increasingly turn to enhanced sampling techniques—such as metadynamics and umbrella sampling [2,3]—and coarse-grained (CG) modeling approaches [4]. These methods enable more comprehensive exploration of the system’s free energy landscape, yet their success critically depends on the selection of appropriate reaction coordinates or collective variables (CVs). CVs must capture the slowest, most functionally relevant motions to accurately reflect thermodynamic and kinetic behavior. However, identifying suitable CVs remains one of the field’s most challenging tasks, typically requiring domain expertise and iterative refinement [5, 6].
\r\nThis complexity has fueled growing interest in machine learning (ML) techniques, which are now transforming how MD simulations are analyzed, interpreted, and even conducted. ML methods have been applied to automate CV discovery, perform dimensionality reduction, build thermodynamic and kinetic models, and enhance sampling efficiency [7]. These models often employ artificial neural networks or graph neural networks to map high-dimensional molecular configurations—such as Cartesian coordinates or molecular descriptors—into low-dimensional representations suitable for analysis [8].
\r\nDepending on the structure and type of data, ML algorithms can be broadly categorized into supervised, unsupervised, and reinforcement learning paradigms [9]. Supervised learning uses labeled input-output pairs to predict properties such as molecular energies or binding affinities [10], while unsupervised learning enables the identification of latent features, such as CVs, directly from data [11].
\r\nA cornerstone of modern ML-driven simulation is the development of symmetry-aware molecular representations. The predictive power of ML models hinges on encoding physical symmetries—like rotation and translation—directly into the model. E(3)-equivariant neural networks have emerged as powerful tools for this purpose, significantly improving data efficiency and generalization in learning potential energy surfaces [12]. Ongoing research continues to explore the optimal balance between enforcing strict symmetry and retaining model flexibility.
\r\nMeanwhile, breakthroughs in structural prediction—most notably the advent of AlphaFold 3—have revolutionized how researchers obtain initial molecular configurations. AlphaFold now provides remarkably accurate models of not only proteins but also their complexes with nucleic acids, ions, and small-molecule ligands [13]. However, these are static snapshots. They cannot capture dynamic behaviors, allosteric transitions, or binding kinetics—areas where physics-based simulations remain indispensable. Initial benchmarks suggest that even state-of-the-art predictors still fall short in modeling protein dynamics and ranking ligand binding affinities, further emphasizing the role of MD [14].
\r\nTo address the dimensionality and sampling bottlenecks, unsupervised ML approaches such as time-lagged autoencoders have reframed CV identification as a data-driven task. More recently, generative models—including diffusion models and variational autoencoders—have emerged as a new frontier. These models can learn the full conformational landscape of biomolecules and enable enhanced sampling, in some cases eliminating the need for predefined CVs altogether [15].
\r\nOnce accurate structural models and CVs are established, ML can significantly improve the estimation of thermodynamic and kinetic properties. In drug discovery, for instance, predicting protein–ligand binding affinity remains a central challenge. ML potentials trained on quantum mechanical data can be combined with enhanced sampling to yield highly accurate free energy landscapes and binding kinetics—results previously unattainable due to computational limitations [16]. However, challenges in data quality, model interpretability, and transferability remain critical areas of ongoing investigation [17].
\r\nFinally, ML is driving a renaissance in CG modeling. Deep neural networks can now learn many-body CG potentials directly from all-atom simulations, capturing emergent properties and enhancing transferability [18]. These models open the door to longer, larger-scale simulations with greater physical accuracy.
\r\nIn this rapidly evolving context, it becomes imperative to critically assess both the promise and limitations of ML in biomolecular simulation. The excitement surrounding these developments must be tempered by careful validation and benchmarking. This workshop thus serves as a timely opportunity—especially for early-career researchers—to explore these cutting-edge methods, engage in constructive dialogue, and chart new directions in the application of machine learning to molecular dynamics and drug discovery.
\r\n
\r\nReferences
\r\n
\r\n[1] J. Behler, M. Parrinello, Phys. Rev. Lett., 98, 146401 (2007)
\r\n[2] P. Sahrmann, G. Voth, Current Opinion in Structural Biology, 90, 102972 (2025)
\r\n[3] K. Kříž, L. Schmidt, A. Andersson, M. Walz, D. van der Spoel, J. Chem. Inf. Model., 63, 412-431 (2023)
\r\n[4] K. Ahmad, A. Rizzi, R. Capelli, D. Mandelli, W. Lyu, P. Carloni, Front. Mol. Biosci., 9, (2022)
\r\n[5] S. Mehdi, Z. Smith, L. Herron, Z. Zou, P. Tiwary, Annual Review of Physical Chemistry, 75, 347-370 (2024)
\r\n[6] H. Zheng, H. Lin, A. Alade, J. Chen, E. Monroy, M. Zhang, J. Wang, AlphaFold3 in Drug Discovery: A Comprehensive Assessment of Capabilities, Limitations, and Applications, 2025
\r\n[7] J. Abramson, J. Adler, J. Dunger, R. Evans, T. Green, A. Pritzel, O. Ronneberger, L. Willmore, A. Ballard, J. Bambrick, S. Bodenstein, D. Evans, C. Hung, M. O’Neill, D. Reiman, K. Tunyasuvunakool, Z. Wu, A. Žemgulytė, E. Arvaniti, C. Beattie, O. Bertolli, A. Bridgland, A. Cherepanov, M. Congreve, A. Cowen-Rivers, A. Cowie, M. Figurnov, F. Fuchs, H. Gladman, R. Jain, Y. Khan, C. Low, K. Perlin, A. Potapenko, P. Savy, S. Singh, A. Stecula, A. Thillaisundaram, C. Tong, S. Yakneen, E. Zhong, M. Zielinski, A. Žídek, V. Bapst, P. Kohli, M. Jaderberg, D. Hassabis, J. Jumper, Nature, 630, 493-500 (2024)
\r\n[8] Fabian B. Fuchs, Daniel E. Worrall, Volker Fischer, Max Welling, NIPS'20: Proceedings of the 34th International Conference on Neural Information Processing Systems, Article No.: 166, Pages 1970 - 1981 (2020)
\r\n[9] H. Sidky, W. Chen, A. Ferguson, Molecular Physics, 118, (2020)
\r\n[10] Y. Wang, J. Lamim Ribeiro, P. Tiwary, Current Opinion in Structural Biology, 61, 139-145 (2020)
\r\n[11] K. Butler, D. Davies, H. Cartwright, O. Isayev, A. Walsh, Nature, 559, 547-555 (2018)
\r\n[12] F. Noé, A. Tkatchenko, K. Müller, C. Clementi, Annu. Rev. Phys. Chem., 71, 361-390 (2020)
\r\n[13] S. Kaptan, I. Vattulainen, Advances in Physics: X, 7, (2022)
\r\n[14] V. Limongelli, WIREs. Comput. Mol. Sci., 10, (2020)
\r\n[15] A. Glielmo, B. Husic, A. Rodriguez, C. Clementi, F. Noé, A. Laio, Chem. Rev., 121, 9722-9758 (2021)
\r\n[16] A. Pak, G. Voth, Current Opinion in Structural Biology, 52, 119-126 (2018)
\r\n[17] M. Lelimousin, V. Limongelli, M. Sansom, J. Am. Chem. Soc., 138, 10611-10622 (2016)
\r\n[18] C. Abrams, G. Bussi, Entropy, 16, 163-199 (2013)\r\n
You can apply to participate and find all the relevant information (speakers, abstracts, program,...) on the event website: https://www.cecam.org/workshop-details/spectrodynamics-2026-connecting-computational-spectroscopic-methods-across-the-electromagnetic-spectrum-1489.
\r\n
\r\nRegistration is required to attend the full event, take part in the social activities and present a poster at the poster session (if any). However, the EPFL community is welcome to attend specific lectures without registration if the topic is of interest to their research. Do not hesitate to contact the CECAM Event Manager if you have any question.
\r\n
\r\nDescription
\r\n
\r\nLight provides one of the most detailed windows into molecules and matter. Modern light sources allow the probing of equilibrium and non-equilibrium phenomena with Å‐level spatial resolution and femto‐ to attosecond temporal precision. Advances in ultrafast laser technology, together with the rise of X-ray free‐electron lasers and next-generation synchrotron sources, have repeatedly pushed the boundaries of spectroscopic methods from low‐frequency collective modes in biomolecules to electronic and core‐level dynamics. An extensive toolbox of linear and multidimensional spectroscopic techniques now spans the entire electromagnetic spectrum. Terahertz (THz) pulses can coherently drive intermolecular and lattice vibrations in solids and soft matter [1], Mid‐IR and Raman methods map vibrational energy (re)distribution in liquids and vibrational signatures of individual modes in complex molecules [2]. Visible spectroscopy tracks ultrafast charge dynamics in chromophores [3] and photochemical molecular pathways [4], while X-ray sources from free-electron lasers and high-harmonic generation setups enabled time-resolved X-ray diffraction of gas‐phase [5] and condensed systems [6].
\r\nDespite sharing common scientific goals, the respective communities have traditionally operated in relative disconnection from each other, relying on different approximations, targeting different observables, and employing distinct numerical implementations. This disconnection manifests, among other symptoms, in the fact that schools, conferences, and workshops are often dedicated to a specific frequency window (e.g. IR spectroscopy) or to simulation methods targeting a class of specific processes (e.g. vibrational dynamics). Opportunities for dialogue and the building of a shared language are lacking. In fact, while preparing this proposal, it became evident that even foundational terms like ab initio or quantum dynamics carry different meanings across communities.
\r\nTo address this fragmentation, the proposed CECAM school brings together researchers from diverse backgrounds to foster mutual understanding and build lasting conceptual bridges. Over five days, participants will engage with both the theoretical foundations and practical implementations of spectroscopies across different communities. We will highlight the fact that despite their apparent differences, all spectroscopic methods can be traced back to a common starting point: a light–matter Hamiltonian that includes the quantum description of electronic, nuclear, and photonic degrees of freedom. From this unified framework, we will explore how different approximations—introduced at various stages—lead to the distinct theoretical approaches adopted in each field.
\r\nThe first part of the school will focus on approaches that solve the exact quantum molecular dynamics in reduced dimensionality. Within this framework, molecules are treated fully quantum-mechanically, while light is treated classically as an external perturbation within the dipole approximation. From the matter perspective, this means that the full electron + nuclear wavefunction is accessible, offering a great level of detail and information, and the accurate treatment of non-Born-Oppenheimer dynamics. From the light perspective, this means that spectroscopic signals are conveniently calculated via the response function approach (RFA) [7], which is however only valid in the weak field limit. Recently, the RFA has been used to design and simulate several spectroscopic signals of femtosecond molecular photochemistry using novel X-ray pulse sources [8], including stimulated X-ray Raman [9], transient X-ray absorption and transmission [10], and many others [11].
\r\nIn the second part, we will shift the focus to longer time scales with more degrees of freedom and study larger molecules in explicit environments (solvent, substrate, etc). In these cases, it is common practice to apply the Born-Oppenheimer approximation and take the classical limit for the nuclei, while keeping the electrons quantum, leading to (finite temperature) molecular dynamics (MD) approaches. To make these simulations computationally tractable, while retaining an explicit description of the electrons, electron–electron interactions are typically simplified using ground-state density functional theory (DFT). This approach, commonly referred to as ab initio molecular dynamics (AIMD), enables the simulation of vibrational spectroscopies such as infrared (IR) and Raman [12,13], as well as surface-specific techniques like sum-frequency generation (SFG) [14,15]. To access larger system sizes and longer simulation timescales, forces can be derived from classical interatomic potentials, facilitating the convergence of multidimensional spectroscopic observables such as THz-Raman spectra [16]. Alternatively, forces can be learned directly from first-principles data using machine-learning (ML) models, enabling ML-driven molecular dynamics and spectroscopy [17-21]. Through path integral techniques, the quantum nature of the nuclei can be recovered, which is particularly important for systems containing light atoms, such as hydrogen [22-24].
\r\nThe third part of the school will explore what happens when the primary interest shifts from vibrational to electronic dynamics. In this context, the electron dynamics at the DFT level can be incorporated by considering its time-dependent version (TDDFT), where the exchange-correlation functionals are usually adiabatic. With this method, UV-visible absorption [25], circular dichroism [26], inelastic X-ray scattering, and electron energy loss [27], and other spectroscopies can be computed. Finally, there are situations in which strong light-matter coupling demands an explicit treatment of the photons [28]. These can be reintroduced either by dressing the Kohn-Sham Hamiltonian with electron-photon exchange-correlation potentials (known as quantum-electrodynamics DFT, or QEDFT) [29] or by a semiclassical treatment of the photons solving Maxwell’s equations (the Maxwell-TDDFT method)[30]. These methods enable the calculation of spectra in cavities or arbitrary electromagnetic environments [31], and can account for polaritonic phenomena, radiative lifetimes, superradiance, and many more.
\r\nThis school brings together leading experts from exact quantum dynamics, ab initio MD, ML‐enabled simulations, and Maxwell–TDDFT to forge a common language and cross‐fertilize ideas. Lectures will cover both the fundamental principles and the latest advances in each area, highlighting current applications and open challenges. Complementing the lectures, hands-on tutorials will reinforce foundational concepts and provide important hands-on experience on several popular computational approaches (see hands-on section below).
\r\nBy spanning the electromagnetic spectrum and the hierarchy of theoretical methods, this school will equip PhD students and postdocs with a unified, multi‐scale, and inter-community perspective on quantum dynamics and spectroscopy. Participants will leave with both a solid grounding in foundational techniques and direct experience of the latest computational frontiers, ready to tackle open challenges in molecular and materials science.
\r\n
\r\nReferences
\r\n
\r\n[1] P. Hamm, The Journal of Chemical Physics, 141, (2014)
\r\n[2] M. Svendsen, K. Thygesen, A. Rubio, J. Flick, J. Chem. Theory Comput., 20, 926-936 (2024)
\r\n[3] F. Bonafé, E. Albar, S. Ohlmann, V. Kosheleva, C. Bustamante, F. Troisi, A. Rubio, H. Appel, Phys. Rev. B, 111, 085114 (2025)
\r\n[4] M. Ruggenthaler, J. Flick, C. Pellegrini, H. Appel, I. Tokatly, A. Rubio, Phys. Rev. A, 90, 012508 (2014)
\r\n[5] J. Flick, D. Welakuh, M. Ruggenthaler, H. Appel, A. Rubio, ACS Photonics, 6, 2757-2778 (2019)
\r\n[6] A. Sakko, A. Rubio, M. Hakala, K. Hämäläinen, The Journal of Chemical Physics, 133, (2010)
\r\n[7] D. Varsano, L. Espinosa-Leal, X. Andrade, M. Marques, R. di Felice, A. Rubio, Phys. Chem. Chem. Phys., 11, 4481 (2009)
\r\n[8] K. Yabana, G. Bertsch, Phys. Rev. B, 54, 4484-4487 (1996)
\r\n[9] Y. Litman, J. Behler, M. Rossi, Faraday Discuss., 221, 526-546 (2020)
\r\n[10] S. Althorpe, Annual Review of Physical Chemistry, 75, 397-420 (2024)
\r\n[11] M. Ceriotti, W. Fang, P. Kusalik, R. McKenzie, A. Michaelides, M. Morales, T. Markland, Chem. Rev., 116, 7529-7550 (2016)
\r\n[12] M. Gastegger, J. Behler, P. Marquetand, Chem. Sci., 8, 6924-6935 (2017)
\r\n[13] R. Han, R. Ketkaew, S. Luber, J. Phys. Chem. A, 126, 801-812 (2022)
\r\n[14] K. Inoue, Y. Litman, D. Wilkins, Y. Nagata, M. Okuno, J. Phys. Chem. Lett., 14, 3063-3068 (2023)
\r\n[15] T. Morawietz, O. Marsalek, S. Pattenaude, L. Streacker, D. Ben-Amotz, T. Markland, J. Phys. Chem. Lett., 9, 851-857 (2018)
\r\n[16] Y. Litman, J. Lan, Y. Nagata, D. Wilkins, J. Phys. Chem. Lett., 14, 8175-8182 (2023)
\r\n[17] D. Nicoletti, A. Cavalleri, Adv. Opt. Photon., 8, 401 (2016)
\r\n[18] T. Ohto, K. Usui, T. Hasegawa, M. Bonn, Y. Nagata, The Journal of Chemical Physics, 143, (2015)
\r\n[19] M. Sulpizi, M. Salanne, M. Sprik, M. Gaigeot, J. Phys. Chem. Lett., 4, 83-87 (2012)
\r\n[20] O. Marsalek, T. Markland, J. Phys. Chem. Lett., 8, 1545-1551 (2017)
\r\n[21] C. Zhang, D. Donadio, F. Gygi, G. Galli, J. Chem. Theory Comput., 7, 1443-1449 (2011)
\r\n[22] D. Keefer, S. Cavaletto, J. Rouxel, M. Garavelli, H. Yong, S. Mukamel, Annu. Rev. Phys. Chem., 74, 73-97 (2023)
\r\n[23] S. Cavaletto, Y. Nam, J. Rouxel, D. Keefer, H. Yong, S. Mukamel, J. Chem. Theory Comput., 19, 2327-2339 (2023)
\r\n[24] D. Keefer, T. Schnappinger, R. de Vivie-Riedle, S. Mukamel, Proc. Natl. Acad. Sci. U.S.A., 117, 24069-24075 (2020)
\r\n[25] M. Kowalewski, B. Fingerhut, K. Dorfman, K. Bennett, S. Mukamel, Chem. Rev., 117, 12165-12226 (2017)
\r\n[26] Shaul Mukamel, Principles of nonlinear optical spectroscopy, Oxford University Press, New York 1995
\r\n[27] J. Kim, S. Nozawa, H. Kim, E. Choi, T. Sato, T. Kim, K. Kim, H. Ki, J. Kim, M. Choi, Y. Lee, J. Heo, K. Oang, K. Ichiyanagi, R. Fukaya, J. Lee, J. Park, I. Eom, S. Chun, S. Kim, M. Kim, T. Katayama, T. Togashi, S. Owada, M. Yabashi, S. Lee, S. Lee, C. Ahn, D. Ahn, J. Moon, S. Choi, J. Kim, T. Joo, J. Kim, S. Adachi, H. Ihee, Nature, 582, 520-524 (2020)
\r\n[28] M. Minitti, J. Budarz, A. Kirrander, J. Robinson, D. Ratner, T. Lane, D. Zhu, J. Glownia, M. Kozina, H. Lemke, M. Sikorski, Y. Feng, S. Nelson, K. Saita, B. Stankus, T. Northey, J. Hastings, P. Weber, Phys. Rev. Lett., 114, 255501 (2015)
\r\n[29] D. Polli, P. Altoè, O. Weingart, K. Spillane, C. Manzoni, D. Brida, G. Tomasello, G. Orlandi, P. Kukura, R. Mathies, M. Garavelli, G. Cerullo, Nature, 467, 440-443 (2010)
\r\n[30] D. Brey, R. Binder, R. Martinazzo, I. Burghardt, Faraday Discuss., 237, 148-167 (2022)
\r\n[31] C. Baiz, B. Błasiak, J. Bredenbeck, M. Cho, J. Choi, S. Corcelli, A. Dijkstra, C. Feng, S. Garrett-Roe, N. Ge, M. Hanson-Heine, J. Hirst, T. Jansen, K. Kwac, K. Kubarych, C. Londergan, H. Maekawa, M. Reppert, S. Saito, S. Roy, J. Skinner, G. Stock, J. Straub, M. Thielges, K. Tominaga, A. Tokmakoff, H. Torii, L. Wang, L. Webb, M. Zanni, Chem. Rev., 120, 7152-7218 (2020)
You can apply to participate and find all the relevant information (speakers, abstracts, program,...) on the event website: https://www.cecam.org/workshop-details/complex-fluids-at-interfaces-structure-stability-and-molecular-effects-1492.
\r\n
\r\nRegistration is required to attend the full event, take part in the social activities and present a poster at the poster session (if any). However, the EPFL community is welcome to attend specific lectures without registration if the topic is of interest to their research. Do not hesitate to contact the CECAM Event Manager if you have any question.
\r\n
\r\nDescription
\r\n
\r\nComplex fluids are ubiquitous in biology, geophysics, and industry [1]. These materials are challenging to characterize and predict [1–4], particularly when they incorporate multiple interfaces, as in colloidal suspensions [4], foams [5–7], or nanoporous membranes [8–10]. Many of these interfaces are micro- or nano-scale and evolve over short times, which can obscure them to observation and pose challenges to experimentalists [2–5, 11, 12]. This opens exciting opportunities for a strong partnership between the development of novel theoretical, computational, and experimental techniques.
\r\nProbing interfaces presents unique challenges compared to probing complex fluids in the bulk. The interfacial structure and constitutive behavior then depend on the composition of two fluids as well as the interfacial configuration [13, 14]. Translating this increased complexity to a computational framework involves developing reliable models describing molecular interactions near fluid-fluid or fluid-solid interfaces [15–17], as well as models for continuum stresses [18]. Molecular modeling is necessary to reveal the physics of chemically-complex structures [17], but is computationally expensive, and it can be challenging to identify the relevant physics to include [19]. Yet the interface also provides unique opportunities for control: in liquid crystals, for example, interfacial stresses can be transmitted through the bulk, leading to novel pattern formation [20] and optical materials exploiting interfacial control [21]. Finally, interfaces are prone to instabilities, which can make flows unpredictable, but opens opportunities to exploit unstable growth for spontaneous patterning.
\r\nTo underscore the present challenges, even for a “simple” Newtonian fluid, the presence of an interface may hinder understanding of flow mechanics. For example, mechanisms for contact during drop impact are still debated [22]: molecular dynamics (MD) simulations can clarify which effects dominate among interfacial instabilities, electrostatic charge, gas-kinetic effects, and other driving forces [22–26], in addition to liquid/surface chemistry [27, 28]. Diffusive processes at interfaces [29] and nanoscale membrane flows, where osmotic and phoretic effects are significant [11, 30], also require further development in MD or coarse-grained models.
\r\n
\r\nThis workshop aims to foster exchanges around the following broad questions:\r\n