Search

Datasets often incorporate various functional patterns related to different aspects or regimes, which are typically not equally present throughout the dataset. We propose a novel, general-purpose partitioning algorithm that utilizes competition between models to detect and separate these functional patterns. This competition is induced by multiple models iteratively submitting their predictions for the dataset, with the best prediction for each data point being rewarded with training on that data point. This reward mechanism amplifies each model's strengths and encourages specialization in different patterns. The specializations can then be translated into a partitioning scheme. The amplification of each model's strengths inverts the active learning paradigm: while active learning typically focuses the training of models on their weaknesses to minimize the number of required training data points, our concept reinforces the strengths of each model, thus specializing them. We validate our concept -- called active partitioning -- with various datasets with clearly distinct functional patterns, such as mechanical stress and strain data in a porous structure. The active partitioning algorithm produces valuable insights into the datasets' structure, which can serve various further applications. As a demonstration of one exemplary usage, we set up modular models consisting of multiple expert models, each learning a single partition, and compare their performance on more than twenty popular regression problems with single models learning all partitions simultaneously. Our results show significant improvements, with up to 54% loss reduction, confirming our partitioning algorithm's utility.

A novel physics-informed operator learning technique based on spectral methods is introduced to model the complex behavior of heterogeneous materials. The Lippmann-Schwinger operator in Fourier space is employed to construct physical constraints with almost no additional overhead, eliminating the need for automatic differentiation. This methodology accelerates the training process by allowing gradient construction on a finite fixed discretization in Fourier space. Later, the spectral physics-informed finite operator learning (SPiFOL) framework is built based on this discretization and trained to map the arbitrary shape of microstructures to their mechanical responses (strain fields) without using any labeled data. The training is done by minimizing equilibrium in Fourier space concerning the macroscopic loading condition, which also guarantees the periodicity. The trained SPiFOL predicts full-field solutions for unseen cases accelerating the conventional fast Fourier transform (FFT) solvers by two orders of magnitude. The accuracy of the SPiFOL is investigated for a wide range of microstructures from dual-phase to multiphase materials and the homogenized stresses computed by SPiFOL show less than 1% error compared to results obtained by conventional FFT solvers. SPiFOL is further enhanced with a Fourier Neural Operator (FNO) to map microstructure to mechanical behavior using only physical equations in Fourier space. Compared to the standard data-driven FNO, SPiFOL shows higher accuracy in predicting stress fields and provides nearly resolution-independent results. Furthermore, the zero-shot-super-resolution feature in heterogeneous domains is investigated. Finally, the effectiveness of SPiFOL as a solver for individual microstructures is demonstrated, highlighting the reduction in solving time when utilizing GPUs.

Soft biological tissues exhibit a tendency to maintain a preferred state of tensile stress, known as tensional homeostasis, which is restored even after external mechanical stimuli. This macroscopic behavior can be described using the theory of kinematic growth, where the deformation gradient is multiplicatively decomposed into an elastic part and a part related to growth and remodeling. Recently, the concept of homeostatic surfaces was introduced to define the state of homeostasis and the evolution equations for inelastic deformations. However, identifying the optimal model and material parameters to accurately capture the macroscopic behavior of inelastic materials can only be accomplished with significant expertise, is often time-consuming, and prone to error, regardless of the specific inelastic phenomenon. To address this challenge, built-in physics machine learning algorithms offer significant potential. In this work, we extend our inelastic Constitutive Artificial Neural Networks (iCANNs) by incorporating kinematic growth and homeostatic surfaces to discover the scalar model equations, namely the Helmholtz free energy and the pseudo potential. The latter describes the state of homeostasis in a smeared sense. We evaluate the ability of the proposed network to learn from experimentally obtained tissue equivalent data at the material point level, assess its predictive accuracy beyond the training regime, and discuss its current limitations when applied at the structural level. Our source code, data, examples, and an implementation of the corresponding material subroutine are made accessible to the public at https://doi.org/10.5281/zenodo.13946282., Comment: 46 pages, 12 figures, 5 tables

The ability to automatically discover interpretable mathematical models from data could forever change how we model soft matter systems. For convex discovery problems with a unique global minimum, model discovery is well-established. It uses a classical top-down approach that first calculates a dense parameter vector, and then sparsifies the vector by gradually removing terms. For non-convex discovery problems with multiple local minima, this strategy is infeasible since the initial parameter vector is generally non-unique. Here we propose a novel bottom-up approach that starts with a sparse single-term vector, and then densifies the vector by systematically adding terms. Along the way, we discover models of gradually increasing complexity, a strategy that we call best-in-class modeling. To identify successful candidate terms, we reverse-engineer a library of sixteen functional building blocks that integrate a century of knowledge in material modeling with recent trends in machine learning and artificial intelligence. Yet, instead of solving the discrete combinatorial problem with 65,536 possible combinations of terms, best-in-class modeling starts with the best one-term model and iteratively repeats adding terms, until the objective function meets a user-defined convergence criterion. Strikingly, we achieve good convergence with only one or two terms. We illustrate the best-in-class one- and two-term models for a variety of soft matter systems including rubber, brain, artificial meat, skin, and arteries. Our discovered models display distinct and unexpected features for each family of materials, and suggest that best-in-class modeling is an efficient, robust, and easy-to-use strategy to discover the mechanical signatures of traditional and unconventional soft materials. We anticipate that our technology will generalize naturally to other classes of natural and man-made soft matter., Comment: 10 pages, 8 figures

Since OpenAI's release of ChatGPT, generative AI has received significant attention across various domains. These AI-based chat systems have the potential to enhance the productivity of knowledge workers in diverse tasks. However, the use of free public services poses a risk of data leakage, as service providers may exploit user input for additional training and optimization without clear boundaries. Even subscription-based alternatives sometimes lack transparency in handling user data. To address these concerns and enable Fraunhofer staff to leverage this technology while ensuring confidentiality, we have designed and developed a customized chat AI called FhGenie (genie being a reference to a helpful spirit). Within few days of its release, thousands of Fraunhofer employees started using this service. As pioneers in implementing such a system, many other organizations have followed suit. Our solution builds upon commercial large language models (LLMs), which we have carefully integrated into our system to meet our specific requirements and compliance constraints, including confidentiality and GDPR. In this paper, we share detailed insights into the architectural considerations, design, implementation, and subsequent updates of FhGenie. Additionally, we discuss challenges, observations, and the core lessons learned from its productive usage.

Sparse regression and feature extraction are the cornerstones of knowledge discovery from massive data. Their goal is to discover interpretable and predictive models that provide simple relationships among scientific variables. While the statistical tools for model discovery are well established in the context of linear regression, their generalization to nonlinear regression in material modeling is highly problem-specific and insufficiently understood. Here we explore the potential of neural networks for automatic model discovery and induce sparsity by a hybrid approach that combines two strategies: regularization and physical constraints. We integrate the concept of Lp regularization for subset selection with constitutive neural networks that leverage our domain knowledge in kinematics and thermodynamics. We train our networks with both, synthetic and real data, and perform several thousand discovery runs to infer common guidelines and trends: L2 regularization or ridge regression is unsuitable for model discovery; L1 regularization or lasso promotes sparsity, but induces strong bias; only L0 regularization allows us to transparently fine-tune the trade-off between interpretability and predictability, simplicity and accuracy, and bias and variance. With these insights, we demonstrate that Lp regularized constitutive neural networks can simultaneously discover both, interpretable models and physically meaningful parameters. We anticipate that our findings will generalize to alternative discovery techniques such as sparse and symbolic regression, and to other domains such as biology, chemistry, or medicine. Our ability to automatically discover material models from data could have tremendous applications in generative material design and open new opportunities to manipulate matter, alter properties of existing materials, and discover new materials with user-defined properties., Comment: 35 pages, 15 figures, 2 tables, 62 references

The constitutive behavior of polymeric materials is often modeled by finite linear viscoelastic (FLV) or quasi-linear viscoelastic (QLV) models. These popular models are simplifications that typically cannot accurately capture the nonlinear viscoelastic behavior of materials. For example, the success of attempts to capture strain rate-dependent behavior has been limited so far. To overcome this problem, we introduce viscoelastic Constitutive Artificial Neural Networks (vCANNs), a novel physics-informed machine learning framework for anisotropic nonlinear viscoelasticity at finite strains. vCANNs rely on the concept of generalized Maxwell models enhanced with nonlinear strain (rate)-dependent properties represented by neural networks. The flexibility of vCANNs enables them to automatically identify accurate and sparse constitutive models of a broad range of materials. To test vCANNs, we trained them on stress-strain data from Polyvinyl Butyral, the electro-active polymers VHB 4910 and 4905, and a biological tissue, the rectus abdominis muscle. Different loading conditions were considered, including relaxation tests, cyclic tension-compression tests, and blast loads. We demonstrate that vCANNs can learn to capture the behavior of all these materials accurately and computationally efficiently without human guidance.

Data-driven methods have changed the way we understand and model materials. However, while providing unmatched flexibility, these methods have limitations such as reduced capacity to extrapolate, overfitting, and violation of physics constraints. Recent developments have led to modeling frameworks that automatically satisfy these requirements. Here we review, extend, and compare three promising data-driven methods: Constitutive Artificial Neural Networks (CANN), Input Convex Neural Networks (ICNN), and Neural Ordinary Differential Equations (NODE). Our formulation expands the strain energy potentials in terms of sums of convex non-decreasing functions of invariants and linear combinations of these. The expansion of the energy is shared across all three methods and guarantees the automatic satisfaction of objectivity and polyconvexity, essential within the context of hyperelasticity. To benchmark the methods, we train them against rubber and skin stress-strain data. All three approaches capture the data almost perfectly, without overfitting, and have some capacity to extrapolate. Interestingly, the methods find different energy functions even though the prediction on the stress data is nearly identical. The most notable differences are observed in the second derivatives, which could impact performance of numerical solvers. On the rich set of data used in these benchmarks, the models show the anticipated trade-off between number of parameters and accuracy. Overall, CANN, ICNN and NODE retain the flexibility and accuracy of other data-driven methods without compromising on the physics. These methods are thus ideal options to model arbitrary hyperelastic material behavior.

For more than 100 years, chemical, physical, and material scientists have proposed competing constitutive models to best characterize the behavior of natural and man-made materials in response to mechanical loading. Now, computer science offers a universal solution: Neural Networks. Neural Networks are powerful function approximators that can learn constitutive relations from large data without any knowledge of the underlying physics. However, classical Neural Networks ignore a century of research in constitutive modeling, violate thermodynamic considerations, and fail to predict the behavior outside the training regime. Here we design a new family of Constitutive Artificial Neural Networks that inherently satisfy common kinematic, thermodynamic, and physic constraints and, at the same time, constrain the design space of admissible functions to create robust approximators, even in the presence of sparse data. We revisit the non-linear field theories of mechanics and reverse-engineer the network input to account for material objectivity, symmetry, and incompressibility; the network output to enforce thermodynamic consistency; the activation functions to implement physically reasonable restrictions; and the network architecture to ensure polyconvexity. We demonstrate that this new class of models is a generalization of the classical neo Hooke, Blatz Ko, Mooney Rivlin, Yeoh, and Demiray models and that the network weights have a clear physical interpretation. When trained with classical benchmark data for rubber, our network autonomously selects the best constitutive model and learns its parameters. Our findings suggests that Constitutive Artificial Neural Networks have the potential to induce a paradigm shift in constitutive modeling, from user-defined model selection to automated model discovery. Our source code, data, and examples are available at https://github.com/LivingMatterLab/CANN., Comment: 31 pages, 14 figures

Understanding real-world dynamical phenomena remains a challenging task. Across various scientific disciplines, machine learning has advanced as the go-to technology to analyze nonlinear dynamical systems, identify patterns in big data, and make decision around them. Neural networks are now consistently used as universal function approximators for data with underlying mechanisms that are incompletely understood or exceedingly complex. However, neural networks alone ignore the fundamental laws of physics and often fail to make plausible predictions. Here we integrate data, physics, and uncertainties by combining neural networks, physics-informed modeling, and Bayesian inference to improve the predictive potential of traditional neural network models. We embed the physical model of a damped harmonic oscillator into a fully-connected feed-forward neural network to explore a simple and illustrative model system, the outbreak dynamics of COVID-19. Our Physics-Informed Neural Networks can seamlessly integrate data and physics, robustly solve forward and inverse problems, and perform well for both interpolation and extrapolation, even for a small amount of noisy and incomplete data. At only minor additional cost, they can self-adaptively learn the weighting between data and physics. Combined with Bayesian Neural Networks, they can serve as priors in a Bayesian Inference, and provide credible intervals for uncertainty quantification. Our study reveals the inherent advantages and disadvantages of Neural Networks, Bayesian Inference, and a combination of both and provides valuable guidelines for model selection. While we have only demonstrated these approaches for the simple model problem of a seasonal endemic infectious disease, we anticipate that the underlying concepts and trends generalize to more complex disease conditions and, more broadly, to a wide variety of nonlinear dynamical systems.

Urban LoRa networks promise to provide a cost-efficient and scalable communication backbone for smart cities. One core challenge in rolling out and operating these networks is radio network planning, i.e., precise predictions about possible new locations and their impact on network coverage. Path loss models aid in this task, but evaluating and comparing different models requires a sufficiently large set of high-quality received packet power samples. In this paper, we report on a corresponding large-scale measurement study covering an urban area of 200km2 over a period of 230 days using sensors deployed on garbage trucks, resulting in more than 112 thousand high-quality samples for received packet power. Using this data, we compare eleven previously proposed path loss models and additionally provide new coefficients for the Log-distance model. Our results reveal that the Log-distance model and other well-known empirical models such as Okumura or Winner+ provide reasonable estimations in an urban environment, and terrain based models such as ITM or ITWOM have no advantages. In addition, we derive estimations for the needed sample size in similar measurement campaigns. To stimulate further research in this direction, we make all our data publicly available., Comment: Author's version of a paper accepted for publication in Proceedings of the 2021 IEEE 94th Vehicular Technology Conference (VTC2021-Fall)