Ferdinando Auricchio
Computational Mechanics and Advanced Materials Group
Department of Civil Engineering and Architecture
University of Pavia, Pavia, Italy
auricchio@unipv.it

TITLE: Additive Manufacturing. A world full of opportunities and challenges!

ABSTRACT: Additive Manufacturing (AM) – also known as 3D printing – is taking off in many industrial processes. In particular, powder bed fusion for metal manufacturing has definitively changed the way of prototyping metal parts but also plastic 3D printing is changing modern engineering in many aspects.
However, AM is a complex physical process, involving different thermo-mechanical phenomena at very different scales; accordingly, simulation is fundamental to predict temperature and stress distributions during and after the printing process. Furthermore, AM allows for new unknown freedom in terms of complex shapes which can be manufactured, opening the door to a new set of design requirements.
The presentation will start with an introduction to AM technologies, possible applications, and will continue with an excursus on our experience on the use of AM to support surgery decision process and industrial developments. Then, the presentation will stear toward more research oriented subjects, focusing on optimization schemes addressing the form freedom characteristic of AM, immersed method to describe the complex physics as well as deposition patterns, lattice components, process uncertainty quantifications. The presentation will close with an excursus on innovative AM technologies under developments in our labs.

SPEAKER: Dennis Kochmann
Dep. of Mechanical and Process Eng.
ETH Zürich

TITLE: Data-Driven Inverse Design of Architected Materials

ABSTRACT: What do periodic truss-based metamaterials, spinodal-type architectures, multi-phase composite materials of general microstructured media have in common? While the forward prediction of their microstructure-dependent effective properties is (relatively) simple, the inverse design of finding suitbale microstructures that ensure certain target effective properties is an open challenge. We will discuss the application of machine learning to tackle this challenge, presenting examples of periodic trusses and spinodoids (using deep neural networks and variational autoencoders) and composite materials (using video diffusion models).

Marco Salvalaglio
Institute of Scientific Computing & DCMS
TU Dresden

TITLE: Modeling pattern formation and analyzing emerging morphologies

ABSTRACT: Pattern formation is ubiquitous in nature. Its modeling and the analysis of emerging morphologies allowed for explaining several physical phenomena, e.g., phase transition and elasticity. Moreover, the design of innovative materials often leverages patterns with peculiar properties. This is the case, for instance, of metamaterials for optical and mechanical applications. In this context, convenient descriptions of patterns are also of central importance, allowing for tailoring structure-property relations and inverse design. This presentation will give an overview of pattern formation, from basic modeling concepts to examples of morphology descriptors. As starting point, renowned models of pattern formation, such as the Cahn Hilliard and Swift-Hohenberg models, will be reviewed, pointing at their basic features and applications. Their analysis will be discussed with convenient representations in Fourier space, a prototypical example of how proper description of data in a broad sense (e.g., field, variables, point clouds) may enable advanced understanding and treatment. Finally, the characterization of patterns with selected descriptors will be discussed, outlining implications for inverse material design.

SHORT BIO: Marco Salvalaglio is an Emmy Noether Group Leader at the Institute of Scientific Computing and the Dresden Center for Computational Materials Science. His research lies at the intersection of materials science, solid-state physics, and applied mathematics. It focuses on both the development of novel mesoscale models and technology-relevant applications. Physicist by training, he obtained his Ph.D. in materials science (and European Doctorate PCAM) at the University of Milano-Bicocca (IT) in 2016. The same year, he joined TU Dresden with a two-year Alexander von Humboldt fellowship and collaborated with the Leibnitz Institute IHP-Microelectronics in Frankfurt(Oder) as a guest scientist. He was invited to the prestigious Hong Kong Institute of Advanced Studies as a Visiting Junior Fellow in 2019. In 2020, he established his research group as a DFG Emmy Noether Programme Grant awardee. The same year he received the Italian habilitation (ASN) as an associate professor in condensed matter physics, and he has been a TU-Dresden Young Investigator since 2021. In 2023, he was nominated MSMSE (IOP) Emerging Leader and was admitted to the Young Academy of Europe.

Prof. Dr. Ivo F. Sbalzarini
Institute of Artificial Intelligence
Faculty of Computer Science, TU Dresden

TITLE: Design Centering and Data-Driven Coarse-Graining

ABSTRACT: We present the fundamental problem of design centering, which aims to find a material design that fulfills given specifications with maximum robustness. We show how a parameterization of meta-materials with mesoscale structure descriptors enables the formulation of inverse design questions as design-centering problems. This enables including non-computable constraints, e.g. for manufacturability or cost, and provides a principled mathematical way of defining the robustness of a design. We also explain a recent machine-learning algorithm, PDE-STRIDE, which can be used to coarse-grain the material properties of a found design to the macroscale from fully-resolved simulation results. This provides interpretable continuum-mechanics models in a data-driven fashion.

SHORT BIO: Ivo F. Sbalzarini graduated in mechanical engineering from ETH Zurich, Switzerland, and holds a doctorate in computer science from ETH Zurich (with Petros Koumoutsakos; Chorafas Award). After intermediate stays at Stanford University, NASA Ames, and Caltech, he became Assistant Professor of Computational Science at ETH Zurich in 2006. Since 2014, Ivo is Professor of Computer Science at Technische Universität Dresden, Germany. He is also a Director of the Center for Systems Biology Dresden (CSBD) and a Senior Research Group Leader with the Max Planck Institute of Molecular Cell Biology and Genetics. His research interests include scientific computing, data-driven modeling, and parallel computing, all with applications in multi-scale problems.

Dr. Bohayra Mortazavi
Leibniz Universität Hannover
bohayra.mortazaviiop@uni-hannover.de

TITLE: First-Principles Multiscale Modeling Enabled by Machine Learning Interatomic Potentials

ABSTRACT: In the conventional finite element (FE) simulations of engineering systems and products, prior to starting the calculations, materials and interactions properties are required to be provided as inputs to the models. Therefore, the development of accurate and robust theoretical approaches for elaborated examinations of various material properties, is highly advantageous in order to enhance the design process and minimize the necessity of complex experimental tests. Since the birth of the concept of machine learning interatomic potentials (MLIPs) in 2007, a growing interest has been developed in the replacement of empirical interatomic potentials (EIPs) with MLIPs, in order to conduct more accurate and reliable molecular dynamics calculations. As an exciting novel progress, recently the applications of MLIPs have been extended towards the exploration of the thermal transport and mechanical responses, providing novel opportunities not heretofore efficiently achievable, neither by EIPs nor by density functional theory (DFT) calculations. In this work, we first illustrate how the combination of EIPs with FE simulations, could provide useful vision on the thermal and mechanical responses of nanostructured materials at the continuum level. Next, by considering several examples of recent studies, the robustness of MLIPs in the analysis of the mechanical and thermal properties will be highlighted, and their advantages over EIPs and DFT methods will be emphasized. It will be finally highlighted that MLIPs furthermore offer astonishing capabilities to marry the robustness of DFT method with continuum mechanics, enabling the first-principles multiscale modeling of mechanical and thermal properties of nanostructures at continuum level, with minimal prior physical knowledge, DFT level of accuracy and affordable computational costs. MLIPs enabled first-principles multiscale modeling is believed to inherit an outstanding prospect to develop fully computerized platforms, to design and optimize novel materials and structures with enhanced performances.

Max Rosenkranz
Institute of Solid Mechanics
TU Dresden

TITLE: Data-driven modeling of viscoelastic materials using physics-augmented neural networks

ABSTRACT: The formulation of suitable constitutive models for novel materials with complex inelastic behavior is still a challenging and time-consuming task. For this reason, many data-driven methods have been developed in recent years. However, many existing methods are not integrated into a physical framework, which comes with several disadvantages. In this work, an artificial neural network based model is presented that strongly satisfies several physical principles and is suitable for viscoelastic materials. Particular attention is paid to thermodynamic consistency as well as compliance with material symmetry. The model is based on the principle of generalized standard materials, which are entirely characterized by two potentials, the free energy and the dissipation potential. Furthermore, a training method is investigated in which the availability of the internal variables is not required. The model is trained and tested with data from linear and non-linear viscoelastic reference models. It is shown that the chosen formulation of the potentials allows for highly accurate predictions for non-linear materials. The mentioned training method yields good results for linear viscoelastic materials. For non-linear materials, however, there are still some challenges to overcome.

Steve WaiChing Sun
Associate professor
Department of Civil Engineering and Engineering Mechanics
Columbia University, New York, USA

TITLE: Non-Euclidean geometric learning for modeling and design of solids

ABSTRACT: This talk presents a framework in which embedding that maps elements between a high-dimensional manifold or graphs and a low-dimensional Euclidean space have been repeatedly used as a remedy to solve three common solid mechanics problems, i.e., (1) formulation of elastoplasticity models, (2) physics-contained digital twins of geometrical nonlinear shells, (3) inverse designs of materials. In the first example, we conceptualize the yielding surface as a hypersurface in a high-dimensional ambient space spanned by stress and internal variables. This setting enables us to precisely represent the onset of yielding and damages for highly complex materials and metamaterials with a collection of coordinate charts collected from RVE simulations. In the second example, our goal is to predict the deformed configurations of Simo-Fox-Rifai shells with an existing database for real-time applications. To generate the response manifold, we introduce a graph isomorphism neural network to embed the finite element solutions onto a latent space, then introduce an over-parametrized neural network to extrapolate the responses. To enforce balance principles, we collect the orthogonal bases from the tangential space of the response manifold and solve the POD problems. As such, there is no need for re-training the neural network while deploying the models for time-sensitive applications. Finally, inverse designs of nonlinear materials often involve finding a level set function that can evolve arbitrarily. To fine-tune a material to exhibit a set of prescribed properties with a principled exploration, we adopt a denoising diffusion probabilistic model where a neural network is trained to reverse the Markov diffusion process from the latent space. By introducing the desired material properties embedded as a context feature vector, the denoising process is guided toward generating microstructures with the designated properties. In all these applications, geometric learning provides us with a cohesive tool to systematically connect isolated results to form a broader perspective for the mechanics and design of materials.

SHORT BIO: Dr. Sun is an associate professor at Columbia University and UPS Foundation visiting professor at Stanford University. He obtained his B.S. from UC Davis (2005); M.S. in civil engineering (geomechanics) from Stanford (2007); M.A. (Civil Engineering) from Princeton (2008); and Ph.D. in theoretical and applied mechanics from Northwestern (2011). Sun’s research focuses on theoretical, computational, and data-driven mechanics for porous and energetic materials. He is the recipient of a few awards, including the Walter Huber Civil Engineering Research Prize (2023), the IACM John Argyris Award (2020), the EMI Leonardo da Vinci Award (2018), the Zienkiewicz Numerical Methods Engineering Prize (2017), and early career awards from NSF, AFOSR, and ARO.

SPEAKER: Gianaurelio Cuniberti
Chair Materials Science and Nanotechnology
Dresden Center for Computational Materials Science (DCMS)
Dresden Center for Intelligent Materials (DCIM)
TU Dresden

TITLE: Mens agitat molem: the age of intelligent materials

ABSTRACT: Materials research is one of the most important technology drivers: new materials essentially contribute to solving current and future social and technical challenges. Frequently, technological leaps and disruptive developments are only made possible by innovations based on novel materials and innovative materials are one of the key technologies for making products and industrial processes economically competitive and sustainable. Modern materials science therefore requires a multidisciplinary approach involving chemistry, physics, engineering and, increasingly, data sciences.
Intelligent materials are programmable and interactive materials that can “feel”, “think” and “act” autonomously without the support of sensors, computers and motors. To explore the full potential of this class of materials, new materials and material-based structures must be discovered and their properties characterized in a wholistic view. Digital research and development processes and the use of data-based and simulative methods make it possible to better predict material properties, making it easier to find the right materials for specific applications.

SHORT BIO: Professor Gianaurelio Cuniberti holds since 2007 the Chair of Materials Science and Nanotechnology at TU Dresden and the Max Bergmann Center of Biomaterials in Dresden, Germany. He studied Physics at the University of Genoa, Italy (where he got his B.Sc. and M.Sc.) and obtained his PhD in 1997 at the age of 27 in a collaboration between the University of Genoa and the University of Hamburg, Germany. He was visiting scientist at MIT and the Max Planck Institute for the Physics of Complex Systems Dresden. From 2003 to 2007, he was the head of a Volkswagen Foundation Research Group at the University of Regensburg, Germany. His research activity is internationally recognized in 491 scientific journal papers to date. He initiated and organized numerous workshops, schools and conferences and took part in international research training networks, offering extensive opportunities for young scientists. He has given plenary and invited talks at numerous international meetings. He received several talent scholarships and awards including the Max Planck Society Schloessmann award (2001) and the VolkswagenStiftung Research Group Individual Grant (2003). He is member of several scientific organizations and corresponding member of the Umbrian Academy of Sciences. Gianaurelio Cuniberti is Honorary Professor at the Division of IT Convergence Engineering of POSTECH, the Pohang University of Science and Technology since 2009, since 2011 Adjunct Professor for the Department of Chemistry at the University of Alabama and since 2019 Guest Professor at SJTU. In 2018 he became faculty member of the transcampus between TU Dresden and King's College London. Professor Gianaurelio Cuniberti is an elected member of the European Academy of Sciences and of the Academia Europaea and of the Germany National Academy of Science and Engineering (acatech).

SPEAKER: Stefano Curtarolo
tbc

TITLE: tbc

ABSTRACT: tbc

SHORT BIO: tbc

Adrian Ehrenhofer
Research Group Leader Materials Informatics
Dresden Center for Intelligent Materials & Institute of Solid Mechanics
Technische Universität Dresden, Germany
adrian.ehrenhofer@tu-dresden.de

TITLE: Towards end-to-end design of active hydrogel composites

ABSTRACT: Active hydrogels play various roles in Soft-Hard Active-Passive Embedded Structures (SHAPES) [1, 2]. As with any active material, the focus of their development lies on the optimization of the obtainable active strain per stimulus change, which can be called "sensitivity" [3]. In this talk, I will discuss how to understand these materials in all stages of the Process-Structure-Property-Performance (PSPP) relationship [4]. I will elaborate, which stages can be (and should be) realized in a data-driven or a white-box way. My talk includes (i) the application of Large Language Models (LLMs) to obtain hydrogel synthesis parameters from scientific literature, (ii) Artificial Neural Network (ANN) based prediction of properties such as the sensitivity, and (iii) the final step of obtaining the SHAPES performance.
[1] Ehrenhofer, A., Design of soft and hard active-passive composite beams. Mechanics of Advanced Materials and Structures, 30(5), 945-960, 2023.
[2] Ehrenhofer, A., Stiffness pairing in soft-hard active-passive actuators, Proc. Appl. Math. Mech., 23, e202200317, 2023
[3] Ehrenhofer, A.; Elstner, M. & Wallmersperger, T. Normalization of Hydrogel Swelling Behavior for Sensoric and Actuatoric Applications, Sensors and Actuators B: Chemical, 2018, 255(2), 1343 - 1353
[4] Wang, Y., Wallmersperger, T. & Ehrenhofer, A., Application of back propagation neural networks and random forest algorithms in material research of hydrogels. Proc. Appl. Math. Mech. 23, e202200278, 2023

SHORT BIO: Dr.-Ing. Adrian Ehrenhofer graduated from Technische Universität Dresden in 2014 and started his PhD thesis with the focus on modeling and simulation of active (smart/intelligent) materials. He worked on the description of Ionic Polymer Metal Composites (IPMC) and permeation through biological membranes using the Poisson-Nernst-Planck multi-field approach. In his PhD thesis, he developed an analogy description for the swelling behavior of active hydrogels. The model was applied for active hydrogel-layered polymeric membranes which are used for microfluidic cell-sorting. He defended his PhD thesis in 2018 and worked as a Post-Doc in the field of smart material modeling for chemo-physical intelligence at Technische Universität Dresden, Germany. Since 2021, he is the Research Group Leader of the Materials Informatics Group at the Dresden Center for Intelligent Materials (DCIM). In 2021, he performed a research stay at the University of Utah, contributing to the multi-field description of Microelectrode Arrays. In 2022, he was the main organizer of the Summer School "Dimensions of Intelligence in Materials" and he co-organizes the current 2023 Summer School "Data-driven exploration and design of materials". He also works as a freelance media creator for tutorial videos in Engineering Mechanics.

Bernardo Ribeiro
LAETA/INEGI - Institute of Science and Innovation in Mechanical and Industrial Engineering, Portugal
Department of Metallurgical and Materials Engineering, Faculty of Engineering of the University of Porto, Portugal
E-Mail: up201806499@fe.up.pt

TITLE: Exploring new RCCAs/RHEAs through High-Throughput screening of the MoNbTaWX (X= Al, Ti, V, Cr, and Zr) system

ABSTRACT: In recent years, Refractory Complexed Concentrated Alloys (RCCAs) and Refractory High Entropy Alloys (RHEAs) have been presented as possible alternatives to the state-of-art Ni-based superalloys due to an outstanding combination of properties under high-temperature service conditions. From several RHEAs, the MoNbTaW system has been mainly explored due to its considerably high yield strength at temperatures. Yet, this alloy presents a brittle behaviour at room temperature, narrowing its applications. As an attempt to improve the MoNbTaW properties, the CALPHAD method for microstructural features prediction, and high-throughput composition screening in the context of developing new RHEA/RCCAs based on the NbMoTaWX (X= Al, Ti, V, Cr, and Zr) system, is presented. Afterwards, the in-situ alloying effect will be assessed at a microstructural and mechanical level by the alloys manufacturing by Direc Energy Deposition (DED).

SPEAKER: Stefano Curtarolo
Edmund T. Pratt, Jr. Distinguished Professor
Materials Science, Electrical Engineering and Physics
Director, Center for Autonomous Materials Design
Duke University, USA - http://materials.duke.edu
Email: Stefano@duke.edu

Techniques to design new ultra-high-temperature-ceramics

ABSTRACT: Disordered multicomponent systems - occupying the mostly uncharted centers of phase diagrams - have been studied for the last two decades for their potential revolutionary properties [1]. Very resilient compositions can be stabilized by maximizing entropy (configurational and/or vibrational) of (near) equimolar mixtures [2]. The search for new systems is mostly performed with trial-and-error techniques, as effective computational discovery is challenged by the immense number of configurations [3]: the synthesizability of high-entropy ceramics is typically assessed using ideal entropy along with the formation enthalpies from density functional theory, with simplified descriptors [4,5] or machine learning methods [6]. With respect to vibrations — even if they may have significant impact on phase stability — their contributions are drastically approximated to reduce the high computational cost, or often avoided with the hope of them being negligible, due to the technical difficulties posed in calculating them for disordered systems [7]. In this presentation I will address many of the problems in the discovery of disordered systems, offer some data-based effective solutions, and discuss the avenues opened by the latter, especially for plasmonic-hyperbolic applications [8]. Research sponsored by DoD-ONR. [1] Nature Reviews Materials 5, 295 (2020); [2] Nature Communications 6, 8485 (2015); [3] npj Comput. Mater. 5, 69 (2019); [4] Nature Communications 9, 4980 (2018); [5] Mater. 159, 364 (2018); [6] Nature Reviews Materials 6, 730-755 (2021); [7] Nature Communications 12, 5747 (2021). [8] Nature Communications 13, 5993 (2022).

SHORT BIO: Prof. Curtarolo is an AI-Materials Scientist. His research interests lie at the intersection of materials science, artificial intelligence, and autonomous discovery of new materials. His current research focuses on theory and discovery of disordered super-hard and ultra-high-temperature ceramics and machine learning approaches to phase stability of alloys. After graduating from MIT, he joined Duke University in 2003. Since then, he has received many national/international awards and recognitions (e.g., ONR Yip, NSF Career, PECASE, IUPAP, Humboldt-Bessel, Highly-Cited 2021 and 2022). At Duke, he directs the Center for Autonomous Materials Design, which has started and currently maintains the AFLOW international data-consortium about materials-information and tools for millions of compounds. The Center and the consortium have also organized several educational [aflow.org/aflow-school/] and outreach initiatives in accelerated materials design [aflow.org/seminars/]. Curtarolo currently leads a MURI team, awarded in 2021, on “SPICES: Spinodal-hardened high-entropy ceramics”.

Prof. Dr.-Ing. habil. Fadi Aldakheel
Institut für Baumechanik und Numerische Mechanik
Leibniz Universität Hannover

TITLE: Efficient multiscale modeling of heterogeneous materials using deep neural networks

ABSTRACT: Material modeling using modern numerical methods accelerates the design process and reduces the costs of developing new products. However, for multiscale modeling of heterogeneous materials, the well-established homogenization techniques remain computationally expensive for high accuracy levels. In this talk, a machine learning approach is proposed as a computationally efficient solution method that is capable of providing a high level of accuracy.

Markus Kästner
Professor for Computational and Experimental Solid Mechanics
Institute of Solid Mechanics
TU Dresden

TITLE: Model- and data-driven methods to explore structure-property linkages

ABSTRACT: The exploration of structure-property linkages of complex microstructures is a challenging task. The difficulties lie in (i) reconstructing plausible 3D statistically representative volume elements (RVEs), e. g., from 2D slices like microscopy images, (ii) modeling the complex and non-linear effective constitutive response and (iii) using it in an efficient multiscale scheme. In this contribution, recently developed methods are presented that aim at addressing these three issues and it is shown how to integrate them into an efficient multiscale workflow. Several examples are shown to demonstrate the effectiveness of the methods.

Department of Civil Engineering and Architecture
University of Pavia, Italy
mihaela.chiappetta@unipv.it

TITLE: Sparse-grids uncertainty quantification of part-scale additive manufacturing processes

ABSTRACT: The present paper aims at applying uncertainty quantification methodologies to process simulations of powder bed fusion of metal. In particular, for a part-scale thermomechanical model of an Inconel 625 super-alloy beam, we study the uncertainties of three process parameters, namely the activation temperature, the powder convection coefficient and the gas convection coefficient. First, we perform a variance-based global sensitivity analysis to study how each uncertain parameter contributes to the variability of the beam displacements. The results allow us to conclude that the gas convection coefficient has little impact and can therefore be fixed to a constant value for subsequent studies. Then, we conduct an inverse uncertainty quantification analysis, based on a Bayesian approach on synthetic displacements data, to quantify the uncertainties of the two remaining parameters, namely the activation temperature and the powder convection coefficient. Finally, we use the results of the inverse uncertainty quantification analysis to perform a data-informed forward uncertainty quantification analysis of the residual strains. \rev{Crucially, we make use of surrogate models based on sparse grids to keep to a minimum the computational burden of every step of the uncertainty quantification analysis. The proposed uncertainty quantification workflow allows us to substantially ease the typical trial-and-error approach used to calibrate powder bed fusion part-scale models, and to greatly reduce uncertainties on the numerical prediction of the residual strains. In particular, we demonstrate the possibility of using displacement measurements to obtain a data-informed probability density function of the residual strains, a quantity much more complex to measure than displacements.

Georgios Tzortzinis
Dresden Center for Intelligent Materials (DCIM),
Technische Universität Dresden
Institute of Lightweight Engineering and Polymer Technology
Technische Universität Dresden

TITLE: 3-D auxetic truss lattices for high performance concrete

ABSTRACT: A study is presented on a new method for confining concrete/mortar materials using steel auxetic truss lattice reinforcement. The study builds on major advances in architected materials and a new concept for the confinement of structural members. Numerical results are presented on a unit cell of the new composite including a mortar matrix and a steel reentrant auxetic lattice as reinforcement, followed by a small scale experimental phase where 3D printed steel auxetic lattices are encased in mortar and tested under axial compression. Several findings emerge: the auxetic lattice applies enhanced confinement to the mortar matrix, the new composite exhibits significant strength increase compared to plain mortar specimens and the composite has a remarkably ductile behavior, with a high residual strength that endures to strains in excess of 20%. The experimental results are further validated through computational modeling. The findings of this paper demonstrate the high potential of using auxetic lattices as reinforcement in a concrete/mortar matrix with advantageous properties for structures in terms of strength and ductility which are crucial in the response of structures under extreme loading conditions such as earthquakes.

Markus J. Buehler
McAfee Professor of Engineering
MIT
mbuehler@mit.edu

TITLE: The mechanics of bio-inspired material intelligence

ABSTRACT: Biological materials offer deep insights into hierarchical design principles, ranging across scales from atoms to structures. In this talk we review how modeling, experiment and synthesis are integrated to understand, design and leverage novel smart material manufacturing for advanced mechanical properties. This allows us to mimic and improve upon natural processes by which materials evolve, and how they meet changing functional needs. We show how we use mechanics to fabricate innovative materials from the molecular scale upwards, with built-in bio-inspired intelligence and novel properties, while sourced from sustainable resources, and breaking the barrier between living and non-living systems. Applied specifically to protein materials, this integrated materiomic approach is revolutionizing the way we design and use materials, and has the potential to impact many industries, as we harness data-driven modeling and manufacturing across domains and applications. The talk will cover several case studies covering distinct scales, from silk, to collagen, to biomineralized materials, as well as applications to food and agriculture, and focuses on mechanistic insights using scaling laws and size effect studies. A specific focus will be on the use of transformer-based attention models as foundational theories, ultimately applied to solve multi-modal material modeling, design and analysis problems.

SHORT BIO: Markus J. Buehler is the McAfee Professor of Engineering at MIT, a member of the Center for Materials Science and Engineering, and the Center for Computational Science and Engineering at the Schwarzman College of Computing. In his research, Professor Buehler pursues new modeling, design and manufacturing approaches for advanced biomaterials that offer greater resilience and a wide range of controllable properties from the nano- to the macroscale. His interests include a variety of functional biomaterial properties including mechanical, optical and biological, linking chemical features, hierarchical and multiscale structures, to performance in the context of physiological, pathological and other extreme conditions. His methods include molecular and multiscale modeling, design, as well as experimental synthesis and characterization. His particular interest lies in the mechanics of complex hierarchical materials with features across scales. An expert in computational materials science and AI, he pioneered the field of materiomics, and demonstrated broad impacts in the study of mechanical properties of complex materials, including predictive materials design and manufacturing. He received many distinguished awards, including the Feynman Prize, the Alfred Noble Prize, the ASME Drucker Medal, the J.R. Rice Medal, and many others. Buehler is a member of the National Academy of Engineering.