Hamidreza Dehghani: Multiscale Poro-Hyperelasticity using ANNs

Hamidreza Dehghani: Multiscale Poro-Hyperelasticity using ANNs

Machine Learning Seminar presentation

Topic: Multiscale Poro-Hyperelasticity using ANNs

Speaker: Hamidreza Dehghani, FSTM, University of Luxembourg

Time: Wednesday, 2021.04.14, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

Poroelastic media are composed of deformable porous solid with viscous fluid percolating its pores. When the solid compartment undergoes large deformation, a hyperelastic model is required (Poro-Hyperelasticity) which is usually carried out via phenomenological homogenized models neglecting complex micro-macro scales interdependency. One reason is that the mathematical two-scale analysis is only straightforward assuming infinitesimal strain theory. Exploiting the potential of ANNs for fast and reliable upscaling and localization procedures, we propose an incremental numerical approach for the analysis of poroelastic media under finite deformation considering infinitesimal strain in each time increment. The Darcy’s experiment is reconstructed numerically and the mechanical response of brain tissue under uniaxial cyclic test is simulated and studied.

Additional material:

https://arxiv.org/abs/2103.06569

Hussein Rappel: Probabilistic modeling and identification of intercorrelated random fields with bounds: Application to linear elastic struts and fibers

Hussein Rappel: Probabilistic modeling and identification of intercorrelated random fields with bounds: Application to linear elastic struts and fibers

Machine Learning Seminar presentation

Topic: Probabilistic modeling and identification of intercorrelated random fields with bounds: Application to linear elastic struts and fibers

Speaker: Hussein Rappel, The Alan Turing Institute

Time: Wednesday, 2021.04.07, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

Many materials and structures consist of numerous slender struts or fibers. Due to the manufacturing process of man-made struts (and the growth process of biological fibers), their mechanical response fluctuates from strut to strut, as well as locally within each strut. In associated mechanical models, each strut is often represented by a string of beam elements. The parameter input fields of each string are ideally such that the local fluctuations and fluctuations between individual strings are accurately captured. The aim of this contribution is to identify the parameters of the random fields from which each set of input fields is considered to be a realization. As only a few sets of input fields are available (due to time constraints of the supposed experimental techniques), this identification problem is ill-posed. A probabilistic identification approach based on Bayes’ theorem is employed to treat this ill-posedness, as well as the involved uncertainties.

Additional material:

Hichem Omrani: Predicting air pollution using remote sensing and sensors measurement

Hichem Omrani: Predicting air pollution using remote sensing and sensors measurement

Machine Learning Seminar presentation

Topic: Predicting air pollution using remote sensing and sensors measurement

Speaker: Hichem Omrani, Luxembourg Institute of Socio-Economic Research (LISER)

Time: Wednesday, 2021.03.31, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

Air pollution is a threat to public health, having negative effects on human health and well-being. This research seminar aims to (1) predict air pollution at a fine spatial resolution using the “land use regression” model, (2) assess population exposure to air population in Luxembourg and its surrounding areas, (3) and discuss ongoing work, opportunities and challenges for future research.

Additional material:

Gabriele Pozzetti: An Industry view on ML: CERATIZIT technology landscape and the art of collaborating with an industrial partner.

Machine Learning Seminar presentation

Topic: An Industry view on ML: CERATIZIT technology landscape and the art of collaborating with an industrial partner.

Speaker: Gabriele Pozzetti, Manager \ Artificial Intelligence and Data Science \ CERATIZIT

Time: Wednesday, 2021.03.24, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

The difference between theory and practice is that in theory the two will mostly agree, in practice they will not.

CERATIZIT is a global player in high-performance material in the midst of its digitalization. ML is nowadays a pervasive technology with plenty of possible applications for an innovative industrial player. But what differentiates a successful project from a failing one? How can a researcher extract value from data instead of using valuable resources to generate data?

This talk aims at giving you an idea of some of the main applications of ML currently active at CERATIZIT and some tips to make your next industrial partnership a success.

Additional material:

 

Damian Mingo Ndiwago: Uncertainty precision and reliability of Ecohydrological models: Bayesian model selection

Machine Learning Seminar presentation

Topic: Uncertainty precision and reliability of Ecohydrological models: Bayesian model selection

Speaker: Damian Mingo Ndiwago, University of Luxembourg, FSTM

Time: Wednesday, 2021.03.17, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

The Bayes factor (BF) is used in Bayesian model comparison and selection. Unlike information-theoretic approaches, it implicitly penalizes the number of parameters in a model. BF can be used for both nested and non-nested models and is invariant to data transformation. Nevertheless, it is sensitive to prior parameter specifications. It may favor a different model for weak prior distributions contrary to the frequentist methods of model selection.

This phenomenon is known as Jeffreys-Lindley’s paradox. BF is undetermined when improper priors are used. However, the pseudo-Bayes (PsBF) is not affected by Jeffreys-Lindley’s paradox. Also, partial Bayes factors such as the Intrinsic Bayes factor (IBF) and the fractional Bayes factor (FBF) are determined for improper priors and are not affected by Lindley’s paradox. Thus, model selection should also report at least the PsBF. If the data set is large, the IBF and FBF should be reported. The IBF and the FBF are less sensitive to outliers.

I will introduce the research and show results based on synthetic data. Then, explain how this will be applied to (Eco)hydrological models with real data.

Additional material:

 

Igor Poltavskyi: Machine learning for molecular simulations

Machine Learning Seminar presentation

Topic: Machine learning for molecular simulations

Speaker: Igor Poltavskyi, University of Luxembourg, FSTM, Department of Physics and Materials Science

Time: Wednesday, 2021.03.10, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

In chemistry and physics, the employment of machine learning (ML) methods has a transformative impact, advancing modeling and improving our understanding of complex molecules and materials. Each ML method comprises a mathematically well-defined procedure, and an increasingly larger number of easy-to-use ML packages for modeling atomistic systems are becoming available. While current approaches mainly focus on developing/improving ML models’ architecture, training sets did not get enough attention. However, training sets are keys to the performance of any ML model, determining its applicability range and predictive power. In this talk, I will address an inherent bias of the reference data caused by their nonuniform nature. On examples of ML force fields trained to reproduce the potential energy surface of molecules, I will demonstrate that the commonly employed measures of the quality of ML models, such as root mean square error, do not provide a full picture. Finally, I will show how combining unsupervised and supervised ML methods can effectively widen the applicability range of ML models to the fullest capabilities of the dataset.

Additional material:

To download the presentation:

Carlotta Montorsi: Tree based algorithms: implementing Conditional Inference Trees and Forest for old age Frailty Index predictions

Machine Learning Seminar presentation

Topic: Tree based algorithms: implementing Conditional Inference Trees and Forest for old age Frailty Index predictions.

Speaker: Carlotta Montorsi, Luxembourg Institute of Socio-Economic Research (LISER)

Time: Wednesday, 2021.02.24, 12:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

Tree-based algorithms are prediction algorithms introduced by Morgan and Sonquist (1963) and popularized by Breiman et al. (1984) almost 20 years later. These algorithms aim at predicting an outcome “out of sample” based on a number of covariates. This is done by partitioning the space of the regressors in non-overlapping regions. When the task is regression, the predicted income is simply the average outcome of units reaching each terminal node. Various methods to grow trees avoiding overfitting exist: Conditional Inference Trees introduced by Hothorn et al. (2006) prevent overfitting by growing the tree conditioning the splitting to a sequence of statistical tests.

In the upcoming presentation, I will present a brief theoretical introduction to the Tree-based model with a particular focus on Conditional Inference Trees and Conditional Inference Forest. Thus, I will show its implementation on real data, namely for predicting a Frailty Index of individuals aged 50+ from different European Countries. Moreover, I will compare the predictive performance of these algorithms with other traditional ML methods and for different subsamples of the training set. Finally, I will discuss the “best predictors” identified by the most accurate of these algorithms in the different subsamples.

Additional material:

[1] Morgan, J. N., and  Sonquist,J. A. (1963). Problems in the Analysis of Survey Data, and a Proposal, Journal of the American Statistical Association, 58(302), 415– 34.
[2] Breiman, L.,Friedman,J.,Stone,C. and R. Olshen (1984). Classification and Regression Trees, Taylor & Francis, Belmont, https://doi.org/10.1201/9781315139470.
[3] Hothorn, T.,Hornik, K. and Zeileis, A. (2006). Unbiased Recursive Partitioning: A Conditional Inference Framework. Journal of Computational and Graphical Statistics.
[4] Brunori, P. and Neidhöfer, G. (2021). The Evolution of Inequality of Opportunity in Germany: A Machine Learning Approach. Review of Income and Wealth.

Minh Vu Chau: Data-driven constitutive laws for hyperelasticity modeling in eigenspace using algorithmic differentiation coupling FEniCS-Pytorch model: application to RVEs

Machine Learning Seminar presentation

Topic: Data-driven constitutive laws for hyperelasticity modeling in eigenspace using algorithmic differentiation coupling FEniCS-Pytorch model: application to RVEs

Speaker: Minh Vu Chau, FSTM, University of Luxembourg

Time: Wednesday, 2021.02.17, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

Macroscale constitutive behavior of heterogeneous hyperelastic materials can be obtained from the microstructure by a homogenization process using multilevel finite elements (FE2) analysis. Although various advancements have been made, FE2 still remains a computationally expensive method. Alternatively, artificial neural networks (ANN) have been proposed as an efficient data-driven method for constitutive modeling, accepting either synthetic data from computational homogenization solutions or experimental datasets. In this presentation, we would like to introduce the Neural Network in which the constitutive relation of hyperelastic materials is achieved via computational homogenization, and is independent upon the used coordinate system.

At the microscopic scale,  a synthetic training dataset is acquired from numerical simulation of RVE using the FEniCS framework. Subsequently, a multivariable regression analysis using ANN is conducted with eigenvalues of strain-stress pairs used as inputs and outputs. After successful training, this ANN is then plugged back into the algorithmic differentiation framework by embedding Pytorch inside FeniCS which creates a coupling symbolic FEniCS-Pytorch model such that it will capture the constitutive relationship of the material. Some technical notes on how to successfully combine multiple sub-losses functions into one total loss are also discussed if we have enough time.

Additional material:

Zhiqiang Zhong: Reinforcement Learning based Meta-path Design for Heterogeneous Graph Neural Network

Machine Learning Seminar presentation

Topic: Reinforcement Learning based Meta-path Design for Heterogeneous Graph Neural Network

Speaker: Zhiqiang Zhong, FSTM, University of Luxembourg

Time: Wednesday, 2021.02.10, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

Heterogeneous Information Networks (HINs), involving a diversity of node types and relation types, are pervasive in many real-world applications. Recently, increasing attention has been paid to heterogeneous graph representation learning (HGRL) which aims to embed rich structural and semantics information in HIN into low-dimensional node representations. To date, most HGRL models rely on manual customization of meta paths to capture the semantics underlying the given HIN. However, the dependency on the handcrafted meta-paths requires rich domain knowledge which is extremely difficult to obtain for complex and semantic rich HINs. Moreover, strictly defined meta-paths will limit the HGRL’s access to more comprehensive information in HINs.

To fully unleash the power of HGRL, we present a novel framework called RL-HGNN, to design different meta-paths for the nodes in a HIN. Specifically, RL-HGNN models the meta-path design process as a Markov Decision Process and uses a policy network to adaptively design a meta-path for each node to learn its effective representations. The policy network is trained with deep reinforcement learning by exploiting the performance of the model on a downstream task. We further propose an extension, RL-HGNN++, to ameliorate the meta-path design procedure and accelerate the training process. Experimental results demonstrate the effectiveness of RL-HGNN and reveal that it can identify meaningful meta-paths that would have been ignored by human knowledge.

Additional material:

Saurabh Deshpande & Milad Zeraatpisheh: Practical implementation session on Neural Networks

Machine Learning Seminar presentation

Topic: Practical implementation session on Neural Networks

Speaker: Saurabh Deshpande & Milad Zeraatpisheh, FSTM, University of Luxembourg

Time: Wednesday, 2021.02.03, 10:00 CET

How to join: Please contact Jakub Lengiewicz

Abstract:

We will create our own neural network from scratch, only using the NumPy library. Afterward, the same example will be implemented in the Keras framework. The session will also involve a Bayesian neural network approach, implemented using the TensorFlow probability library.

Additional material:

Shared code: https://colab.research.google.com/drive/1aTf490wwL7r8_eEjcW4Sw9rgPg6ntVvf#scrollTo=B1OA9W_KnmUo