Machine Learning for Simulation

Project Network 6

Interplay of the RQs and ML

So far, classical simulations and Machine Learning (ML) are separated fields based on different paradigms. This Project Network paves the way for integrating them with joint model-based and data-driven predictions capable of tackling simulation problems that are currently out of reach. In addition, opening the black box of ML through visualization will give non-experts access to these tools. On the one hand, the network will focus methodologically on supervised, unsupervised, and reinforcement learning. We will develop, analyze, and implement novel ML strategies adapted to the characteristics of simulation data, such as symmetries, invariances, and physical prior knowledge. And, surprisingly, even deterministic trajectories are incompatible with statistical assumptions required for standard learning guarantees for ML methods. The new ML techniques will further support visual exploration and interactive control by users and will be applied to the different stages of the simulation cycle, specifically in the modeling, prediction, visualization, and visual analytics stages. On the other hand, and complementary to these methodological considerations, PN 6 will aim at broad collaborations with the other networks by disseminating our network’s ML and visualization competences as well as tools to the diverse application fields. To this end, we will combine ML expertise in mathematics and computer science in this network with domain expertise from collaborating networks.

Research Questions

RQ 1 ML4Sim: How can ML be adopted for simulation; specifically, how can we incorporate application-specific physical constraints and prior knowledge, and how can we learn in the absence of ML-specific statistical assumptions?

RQ 2 Sim4ML: How can simulation data augment real-world/experimental data for ML?

RQ 3 VisML2: How can we use visualization to better understand and control ML and, vice versa, how can we leverage ML for data-integrated visualization?

Project Overview

PN 6-1 Deep greedy kernel methods for submodel coupling in fluid- and biomechanics
PN 6-1 (II) Greedy deep kernel methods for data-based-modelling in Biomechanics
PN 6-2 Surrogate Modelling by Simulation-Enhanced Machine-Learning
PN 6-2 (II) Towards Parameter-Dependent Data-Enriched Physics-Informed Machine Learning
PN 6-3 Understanding Physical Constraints in Machine Learning for Simulation
PN 6-3 (II) Gaussian Process Techniques for Differential Equations
PN 6-4 Visual Analytics for Deep Learning
PN 6-4 (II) Visual Analytics for Machine Learning
PN 6-5 Data-integrated Simulation of Human Perception and Cognition
PN 6-5 (II) Interpretable and explainable cognitive inspired machine learning systems
PN 6-6 Machine Learning for Data-Driven Visualization (ML4Vis)
PN 6-7 Machine Learning for Bayesian Model Building (M4LBMB)
PN 6-8 Visual Data Science to Master Complex Simulation Ensembles
PN 6-9 Analytical models combined with data-driven techniques in dynamic Compton scattering tomography
PN 6-10 Generalization and Robustness of Learned Simulators (GRLSim)
PN 6-11 Data-Integrated Simulation of Interactive Behaviour
PN 6-12 Simulation-Based Prior Distributions for Bayesian Models (SBPriors)

Associated Projects

Project Network Coordinators

This image shows Dirk Pflüger

Dirk Pflüger

Prof. Dr. rer. nat.

[Photo: SimTech/Max Kovalenko]

This image shows Ingo Steinwart

Ingo Steinwart

Univ.-Prof. Dr. rer. nat.

[Photo: SimTech/Max Kovalenko]

Daniel Weiskopf

Prof. Dr.
To the top of the page