Schedule

The [ML-DE] Workshop features a packed schedule of keynotes, sessions, and networking opportunities.

Workshop Date

Sunday, 26th of October 2025, Full Day

Location

ECAI 2025, Bologna, Italy at University of Bologna - The Engineering School

Conference Timetable

9:00-9:05 Opening Remarks

Chair: Cecília Coelho and Bernd Zimmering

9:05-9:50 Keynote by Andrei A. Klishin

Chair: Bernd Zimmering
Title: Statistical Mechanics of Dynamical System Identification

9:50-10:30 Session 1: Foundations — Learning Dynamics & Operators

Chair: Michael Hohmann

  • 9:50–10:10 A Unified Framework for Neural Computation and Learning Over Time
  • 10:10–10:30 ExtremONet: Extreme-Learning-based Neural Operator for identifying dynamical systems.

10:30-11:00 Coffee Break

11:00-13:00 Session 2: Modeling Physical Systems with ML

Chair: Luís Ferrás

  • 11:00–11:20 Learning non-Markovian Dynamical Systems with Signature-based Encoders
  • 11:20–11:40 Breaking Free: Decoupling Forced Systems with Laplace Neural Networks
  • 11:40–12:00 TaylorNet: Learning PDEs from Non-Grid Data
  • 12:00–12:20 Hamiltonian Normalizing Flows as kinetic PDE solvers: application to the 1D Vlasov-Poisson Equations
  • 12:20–12:30 Wrap-Up & Afternoon Outlook

12:30-14:00 Lunch Break

14:00-14:45 Keynote by Ricardo Baptista

Chair: Cecília Coelho
Title: Memorization and Regularization in Generative Diffusion Models

14:45-15:30 Session 3: Sequences, Memory & Generative Models

Chair: Niklas Kompe

  • 14:45–15:05 Physics-Informed Graph Neural Networks for Air Pollution Forecasting in the Netherlands
  • 15:05–15:25 Flowing Straighter with Conditional Flow Matching for Accurate Speech Enhancement
  • 15:25–15:30 Poster Session Preparation

15:30-16:00 Coffee Break

16:00-17:30 Session 4 & Poster Session — Applications Lightning Talks

Chair: Oliver Niggemann

  • 16:00–16:10 Digital Twin, Machine Learning and Differential Equations: Application to the population dynamics of an invasive species
  • 16:10–16:20 An Informed Autoencoder-Architecture for the Estimation of Mean Muscle Parameters of the Myoelectric Potential Field
  • 16:20–16:30 An Hybrid Neural Network and Optimal Control Approach for Irrigation Scheduling
  • 16:30–16:40 A Deep Learning Powered Numerical Relativity Surrogate for Binary Black Hole
  • 16:40–16:50 Closing Remarks
  • 16:50–17:30 Poster Session