Detailed Program of the event
The meeting will feature invited keynote lectures, contributed presentations, and flash (10 min) talks within a single-track format to encourage discussion and exchange.
This is an in-person only event.
The program and events details can be found in this PDF document.
Program
Day 1, 28 April 2026
9:00-9:30 Welcome to the 662 colloquium – opening AC
9:30 -10:20 Keynote – Chair EC
(50 minutes, 40 minutes talk, 10 minutes questions)
Closing the gap between simulations and reality
Nathan Kutz, University of Washington
The requirements of modern sensing are rapidly evolving, driven by increasing demands for data efficiency, real-time processing, and deployment under limited sensing coverage. Complex physical systems are often characterized through the integration of a limited number of point sensors in combination with scientific computations which approximate the dominant, full-state dynamics. Simulation models, however, inevitably neglect small-scale or hidden processes, are sensitive to perturbations, or oversimplify parameter correlations, leading to reconstructions that often diverge from the reality measured by sensors. This creates a critical need for data assimilation, the process of integrating observational data with predictive simulation models to produce coherent and accurate estimates of the full state of complex physical systems. We propose a machine learning framework for Data Assimilation with a SHallow REcurrent Decoder (DA-SHRED) which bridges the simulation-to-real (SIM2REAL) gap between computational modeling and experimental sensor data. For real-world physics systems modeling high-dimensional spatiotemporal fields, where the full state cannot be directly observed and must be inferred from sparse sensor measurements, we leverage the latent space learned from a reduced simulation model via SHRED, and update these latent variables using real sensor data to accurately reconstruct the full system state. Furthermore, our algorithm incorporates a sparse identification of nonlinear dynamics based regression model in the latent space to identify functionals corresponding to missing dynamics in the simulation model. We demonstrate that DA-SHRED successfully closes the SIM2REAL gap and additionally recovers missing dynamics in highly complex systems, demonstrating that the combination of efficient temporal encoding and physics-informed correction enables robust data assimilation.
10:20-10:50 Coffee break
10:50-12:30 Standard talks (5) – Chair AF
(20 minutes:
15 minutes presentation, 4 minutes questions, 1 minute for change over)
10:50 – 11:10
Tim .J. Rogers, University of Sheffield
Operational Identification of Nonlinear Systems in Steady-State
T.J. Rogers, M.D. Champneys, K. Worden
11:10 – 11:30
Davide Murari, University of Cambridge
Discovering Piecewise-Smooth Dynamics from Trajectory Data
D. Murari, C. Budd, E. Jansson, C.-B. Schönlieb
11:30 – 11:50
Jan Decuyper, Vrije Universiteit Brussel
Physics-informed identification of nonlinear systems via dynamical invariants
J. Decuyper, G. G. Cruz, M. Schoukens, T. De Troyer, M. C. Runacres
11:50 – 12:10
Alessandra Vizzaccaro, Politecnico di Milano
Reduced order modelling of Hopf bifurcations for the Navier-Stokes equations through invariant manifolds
A. Colombo, A. Vizzaccaro, C. Touzé, A. Stabile, L. Pastur, A. Frangi
12:10 - 12:30
Romit Maulik, Purdue University
Nudging-based data assimilation for interpretable scientific machine learning
Hojin Kim , Ashwin Suriyanarayanan, Melissa Adrian, Romit Maulik
12:30-14:00 Lunch (includes coffee)
14:00-15:00 Flash Talks (6) – Chair AM
(10 minutes slot,
9 minutes presentations, no questions, with one minute changeover)
14:00 – 14:10
Tobias Weidemann, University of Stuttgart
Estimating Amplitude Bounds of Chaotic Strongly Modulated Vibrations Based on Limited Data
T. Weidemann, A. Cicirello, M. Krack
14:10 – 14:20
Kacper Cerek, Hamburg University of Technology
Data-Driven Surrogate Modelling of Quay Walls Utilizing Hybrid Neural Network Architectures
K. Cerek, F. Williams-Riquer, M.A. Abdennadher, E. Hadjiloo, J. Grabe
14:20 – 14:30
Alan Xavier, Imperial College London
Neural Operators For Accelerating Flow Field Predictions in a 2D Compressor Cascade
A.Xavier, L.Renson
14:30 – 14:40
Mario Sinani, University of Washington, Imperial College London
Physics-Informed Neural ODEs for Nonlinear Parametric Aeroelastic Modeling
M. A. Sinani, B. Moseley, A. Wynn, S. L. Brunton
14:40 - 14:50
Nicola Farenga, Politecnico di Milano
Learning the continuous-time dynamics: from trajectories to velocities
N. Farenga, A. Manzoni
14:50 - 15:00
Paul Fährmann, TU Dortmund
On the Koopman Autoencoder for high-dimensional liquid metal flow dynamics
Paul Fährmann, Michael Skowronek, Andreu Queralt McBride, Dmitry Krasnov, Jörg Schumacher, Sebastian Peitz
15:00-15:30 Coffee break
15:30-16:30 Talks by Early Career Awardee Nominations (3) – Chair PB
(20 minutes:
15 minutes presentation, 4 minutes questions, 1 minute for change over)
15:30 – 15:50
Matteo Rufolo, IDSIA USI-SUPSI, Switzerland
Meta-Symbolic Regression of Physical Parameters in a 2-DOF Planar Robot Arm
M. Rufolo, D. Piga, M. Forgione
15:50 – 16:10
Nicolò Botteghi, Politecnico di Milano
Robust Control of Parametrized Nonlinear Dynamical Systems with Hypernetwork-based Reinforcement Learning
Nicolò Botteghi, Gabriele Pascali, Matteo Tomasetto, Urban Fasel, Mengwu Guo
16:10 – 16:30
Yi Luo, Leibniz University Hannover
Dimension-reduced probability density evolution equation for vehicle- rail-bridge systems
Yi Luo , Peng Yuan, Michael Beer
16:30-17:00 Spotlight talk - chair AC
(30 minutes, 25 minutes talk, 5 minutes questions)
Keith Worden, University of Sheffield
On Bayesian Tree-Adjoining Grammars for Grey-Box System Identification
C.A. Lindley & K. Worden
17:00-18:00 Break
Welcome apéro at Villa Grumello: 18.00 – 21:00
Day 2, 29 April 2026
9:00-9:30 Early Career Invited talk 1– Chair PB
(30 minutes, 20 minutes talk, 10 minutes questions)
Georgios Kissas, ETH Zurich
Verifiable Scientific Discovery with Formal Grammars
Georgios Kissas
9:30 -10:20 Keynote – Chair AM
(50 minutes, 40 minutes talk, 10 minutes questions)
Probabilistic regime dependent physics-informed machine learning models for structural assessment
Lizzy Cross, University of Sheffield
How much we know about our structures and how they respond to different environments and operating conditions is variable, as too is our confidence in that knowledge. Embedding physics into machine learning algorithms can be helpful in many ways, but not when what we know doesn't generalise across our data and indeed operational envelope. This talk will explore how we can embed differing physical insights into regime dependent and aware algorithms also taking into account variable uncertainty on our assumptions and data.
10:20-10:50 Coffee break
10:50-12:10 Standard talks (4) – Chair EC
(20 minutes:
15 minutes presentation, 4 minutes questions, 1 minute for change over)
10:50 – 11:10
K. Vlachas, ETH Zurich
Weight-Space Interpolation for Parametric Dynamical Systems via Hypernetworks with Learned Embeddings
P. R. Vlachas, K. Vlachas, E. Chatzi
11:10 – 11:30
Mengwu Guo, Lund University
Principled Gaussian Process Modeling for Differential Equations
Mengwu Guo
11:30 – 11:50
Tore Butlin, University of Cambridge
Gaussian process regression for efficient equivalent linearisation of nonlinear systems
J. Hickey, T. Butlin
11:50 – 12:10
József Kövecses, McGill University
High-Fidelity Interaction Force Modelling in Mechanical Systems: A Hybrid Analytical–Data-Driven Approach
Eric Karpman, Mahdi Maleki, József Kövecses
12:10 -13:30 Lunch (includes coffee)
13:30-14:30 Flash Talks (6) – Chair AM
(10 minutes slot,
9 minutes presentations, no questions, with one minute changeover)
13:30 – 13:40
Giorgio Luigi Morales Luna, Aston University
Discovering Non-Linear Equations Under Epistemic Uncertainty Using Transformer- Based Multi-Set Skeleton Prediction
Giorgio Morales, John W. Sheppard
13:40 – 13:50
Teng Ma, Politecnico di Milano
Encoding cumulation to learn perturbative nonlinear oscillatory dynamics
Teng Ma, Luca Rosafalco, Wei Cui, Lin Zhao, Attilio Frangi
13:50 – 14:00
Sarvin Moradi, Technical University of Eindhoven
Energy-Based Port-Hamiltonian Neural Networks for Learning Nonlinear Dissipative Dynamical Systems
S. Moradi, N. Jaensson, R. Tóth, M. Schoukens
14:00 – 14:10
Matteo Tomasetto, Politecnico di Milano
Physics-enhanced reinforcement learning for real-time optimal control of dynamical systems
Matteo Tomasetto, Nicolò Botteghi, Gabriele Bruni, Andrea Manzoni
14:10 - 14:20
Tairan Wang, University of Southampton
From Bayesian sampling to data-driven inference in model updating: a generative AI perspective
Tairan Wang, Sifeng Bi
14:20 - 14:30
Francisco Williams-Riquer, Hamburg University of Technology
Dynamic Mode Decomposition with Time-Delay Embeddings for the Analysis of Nonlinear Soil Response to Vibrodriving-Induced Vibrations
F. Williams-Riquer, M.A. Abdennadher, K. Cerek, J. Grabe
14:30-15:30 Talks by Early Career Awardee Nominations (3) – Chair PB
(20 minutes:
15 minutes presentation, 4 minutes questions, 1 minute for change over)
14:30 – 14:50
Luca Rosafalco, Politecnico di Milano
NF-SINDy for discovering partially or indirectly observed microsystems
Luca Rosafalco , Alessio Colombo and Attilio Frangi
14:50 – 15:10
Meng-Ze Lyu, Leibniz University Hannover
Physics-Driven Dimension-Reduced Probability Density Evolution for Data-Driven High-Dimensional Nonlinear Stochastic Dynamics
Meng-Ze Lyu, Jian-Bing Chen, Michael Beer
15:10 – 15:30
Matteo Torzoni, Politecnico di Milano
Learning through action in mechanical systems via free energy minimization
M. Torzoni, D. Maisto, A. Manzoni, F. Donnarumma, G. Pezzulo, A. Corigliano
Break: 15:30-16:00
Social activities from 4pm:
The Lake Como funicular Or The Lake Como boat tour
Day 3: 30 April 2026
9:00-9:30 Early Career Invited talk 2– Chair EC
(30 minutes, 20 minutes talk, 10 minutes questions)
Nicola Rares Franco, Politecnico di Milano
Learning adaptive basis representations for parametrized PDEs with Deep Orthogonal Decomposition
Nicola Rares Franco, Andrea Manzoni, Paolo Zunino, Jan S. Hesthaven
9:30-10:30 Standard talks (3) – Chair AF
(20 minutes:
15 minutes presentation, 4 minutes questions, 1 minute for change over)
09:30 – 09:50
Max Champneys, University of Sheffield
Bayes-by-backpropagation through neural ordinary differential equations for nonlinear system identification
M. D. Champneys, M. R. Jones, T. J. Rogers
09:50 – 10:10
Ioannis Koune, Delft University of Technology
Scalable Variational Inference for Systems with Complex Hierarchical Dependencies
I. C. Koune, A. Cicirello
10:10 – 10:30
Pengyu Zhang, University of Cambridge
Population-level Bayesian Inference and Neural Closure Learning with Jointly Trained Surrogates
Pengyu Zhang, Arnaud Vadeboncoeur, Alex Glyn-Davies, Mark Girolami
10:30-11:00 Coffee break
11:00-12:00 Standard talks (3) – Chair AM
(20 minutes:
15 minutes presentation, 4 minutes questions, 1 minute for change over)
11:00 – 11:20
Andrew Bickerdike, University of Exeter
AI-Enhanced Vibrational Capsule for Minimally Invasive Detection of Abnormal Bowel Tissue
Y. Liu
11:20 – 11:40
Francesco Clementi, Università Politecnica delle Marche
Comparative Machine Learning Techniques for Predicting the Fundamental Frequency of Masonry Towers from a Large Experimental Database
S. Mochetti, F. Roscini, G. Misseri, M. Betti, F. Clementi
11:40 – 12:00
Konstantinos Karapiperis, EPFL
Physics-informed learning of transient granular rheology
K. Karapiperis, H. Menon, G. Anantha Padmanabha
12:00 -13:00 Lunch (includes coffee)
13:00 -14:00 Panel Discussion – Colloquium organisers
“Future of Physics-enhanced machine learning and data-driven nonlinear dynamics”
14:00-16:00 Closing and open discussions