Spring School Speakers

Alberto Pérez de Alba Ortíz

Alberto Pérez de Alba Ortíz portrait
Alberto Pérez de Alba Ortíz

Wednesday 10:30 - 11:30  Dr. Alberto Pérez de Alba Ortíz

Introduction to AI and ML, basics, real life and research examples

Material and molecular design in the age of AI: In living organisms, a fascinating range of behaviors are regulated by chemical, conformational and phase transitions with evolutionarily-tuned free-energy barriers. Designing such transitions in synthetic soft materials is key to achieve living-like, responsive technologies for health and sustainability. But how can we dissect and control such complex processes at the nanoscale? After a general introduction to ML approaches, I will show past and present work in developing and applying forefront computational simulation and ML methods—often in combination with experiments—to understand and design mechanistic pathways for a variety of systems. In particular, I will discuss transitions in several biopolymers—e.g, polysaccharides, peptides, signaling proteins and DNA—as well as in colloidal systems, which have been resolved via enhanced sampling, multiscale modeling, and a variety of supervised and unsupervised ML approaches. I will conclude with future perspectives to advance our insight and control of matter at the nanoscale.

Biography: Alberto is currently an assistant professor in Computational Soft Matter at the University of Amsterdam, where he works on understanding and designing molecules and materials via simulations and machine learning. Before, he was a postdoctoral researcher in Prof. M. Dijkstra’s Soft Condensed Matter group at Utrecht University, working on inverse design of self-assembly. Previously, he completed his PhD, cum laude, in Dr. B. Ensing’s Computational Chemistry group at the University of Amsterdam, where he developed and applied path-based free-energy methods for biomolecular transitions. He obtained his MSc in Computational Science and Engineering at the Technical University of Munich, where he wrote his thesis on adaptive quantum mechanics/molecular mechanics simulations at Prof. Karsten Reuter’s Theoretical Chemistry group.

Jana M. Weber

Jana Weber portrait
Jana M. Weber

Wednesday 11:30 - 12:30  

Machine Learning, Catalysts, Energy Materials, Methods, Models: Molecular machine learning for property prediction and molecular design. 

The discovery and development of novel functional molecules impacts the sustainability transition through a multitude of aspects, e.g. by discovering improved (bio)catalysts, novel energy materials, biodegradable products, or more sustainable reaction pathways. Communities from computational chemistry, material design, and bioinformatics are jointly supporting such endeavours through the development of machine learning algorithms on molecular data: molecular machine learning. In this talk, we will see recent computational concepts such as self-supervised learning, language models, and generative AI applied on molecular data. I will highlight selected building blocks for molecular design, such as finding a suitable molecular representation, developing molecular property predictors, and strategies for inverse molecular design.

Biography: Jana M. Weber is an assistant professor for artificial intelligence in bioscience at TU Delft in the Department of Intelligent Systems since 2022. There, she manages the AI4b.io lab, and she is part of the Delft Bioinformatics Lab. In 2022, Jana also defended her PhD in Chemical Engineering from the University of Cambridge for her work on circular chemistry through large scale reaction network analysis and optimization. Prior to that, she obtained her MSc in Environmental (Process) Engineering from RWTH Aachen University, with her Master’s thesis performed at the Jülich Research centre at the institute of bio- and geosciences. Jana’s research interests cover molecular machine learning and network science for a broad range of environmental applications.

Robert Pollice

R. Pollice portrait
Robert Pollice

Thursday 09:00 - 10:00  

Artificial Organic Chemistry

In this lecture, I will explain some of the applications of artificial intelligence in organic chemistry, with a particular focus on the design and synthesis of molecules.

For that, we will look into the three main branches of machine learning and explore the mathematical foundation of some of the most popular artificial neural networks. We will cover some of the most important representations of molecules in the context of machine learning, such as SMILES and SELFIES. Based on that foundation, I will introduce some of the algorithms we have developed that can be used for inverse molecular design. Finally, I will talk about the key concepts behind Artificial Intelligence approaches for reaction prediction and computational retro-synthesis.

Biography: In 2015, Robert received a Master of Science in Technical Chemistry, working with Professor Michael Schnürch on mechanistic investigations of C-H activation reactions. Subsequently, he moved to Switzerland to join the group of Professor Peter Chen at ETH Zurich for his doctoral studies investigating London dispersion in molecular systems. Afterward, he decided to work in the group of Professor Alán Aspuru-Guzik at the University of Toronto as a postdoctoral fellow to conduct research on organic electronic materials, lab automation and machine learning. In August 2022, Robert started as Assistant Professor for Computer-Aided Organic Synthesis at the University of Groningen.

Marc-Etienne Moret

M-E. Moret portrait
Marc-Etienne Moret

Thursday 10:00 - 11:00

Computational chemistry in homogenous catalysis

Computational chemistry, in particular Density Functional Theory (DFT), has become an essential tool for the study of catalytic reactions. DFT calculations can provide detailed insight into reaction mechanisms and guide the design of new, efficient catalysts. In this lecture, I will discuss the use of DFT calculations as a “companion” for experimental researchers in molecular chemistry. After a concise introduction to the theory (what are we actually calculating ?) and some practical guidelines, I will discuss several recent examples illustrating how theory and experiments can work together to further our understanding of organometallic and catalytic reactions.

Biography: After an M.Sc. Thesis in computational chemistry at EPFL (with Prof. Ursula Rothlisberger), Marc-Etienne opted for experimental work for his PhD (2009) at ETHZ (with Prof Peter Chen). There, he pioneered the study of unsupported Pt–Cu bonds using a combination of solution-phase, mass-spectrometric and computational methods. In his postdoctoral work at CalTech (with Prof. Jonas C. Peters), he developed Fe complexes of donor/acceptor tripodal ligands that ultimately led to the first Fe-based homogeneous catalytic reduction of N2. He moved to Utrecht University in 2012 to start his independent academic career and is now Associate Professor in the Organic Chemistry and Catalysis group. His research group, funded amongst others by an ERC Starting Grant (2017) and several NWO grants, investigates the design of cooperative ligands for organometallic and catalytic reactions using earth-abundant elements.

Zeila Zanolli

Picture of dr. Zeila Zanolli
Zeila Zanolli

Thursday 11:00 - 12:00  

Quantum Materials by Design

Major advances in human civilization are driven by developments in materials. This is such a remarkable feature that historical eras are named after the material (and the related technology) that dominated that time. Today we live in the silicon era: Silicon technology enables our modern way of life via mobile phones, computers, automation. However, we are reaching the limits of silicon technology, as the related energy demand is not sustainable. It is time to move forward. As a scientist, we work to answer the question: what is the material that will enable the next revolution?

In this lecture, I will introduce quantum materials as a possible way forward. These are materials where quantum behavior is visible at macroscopic scales and, hence, can be used as a basis for developing quantum technologies.  Using the predictive power of first-principles techniques, I will explain how to understand, control and design quantum materials and how to exploit the quantum effects emerging at reduced dimensionality and at interfaces to achieve dissipationless (topological) carrier transport, tune charge and spin relaxation processes, or identify spectroscopic signatures of defects in 2D layered materials.

Biography: Zeila Zanolli is Associate Professor in Theory and Simulation of Quantum Materials at Utrecht University since 2020. She has over 20 years of experience in the study of quantum materials (Dirac materials, topological insulators, superconductors) using Density Functional Theory (DFT) to predict electronic, magnetic and topological properties, many-body techniques to address non-equilibrium quantum electron transport (Non-Equilibrium Green’s Function + DFT) and theoretical spectroscopy (GW approximation/Bethe-Salpeter Equation). She was among the first to demonstrate that magnetism and topological features can be controlled by proximity interaction at interfaces. Her group is among the developers of the SIESTA and YAMBO codes. In her career, she organized numerous international schools on Quantum Materials (EPS, U. Pisa), first-principles simulations (CECAM, MaX), and related conferences (EPS, ETSF). She serves in the board of the European Theoretical Spectroscopy Facility (2018 - present) and of the ”Semiconductors and Quantum Materials” section of the European Physical Society (2022 – present). She was elected fellow (2017) and board member (treasurer 2018-22) of the Young Academy of Europe.

Gerardo Campos Villa Lobos

Gerardo Campos Villa Lobos portrait
Firstname Surname 2

Thursday 17:00 - 18:00 Gerardo Campos Villa Lobos

Utilizing Machine Learning for the Bottom-Up Coarse-Graining of Colloidal Systems

Colloids consist of mesoscopic particles dispersed in a fluid and are maintained in suspension by thermal fluctuations. These systems, which historically played a pivotal role in unraveling the discontinuous nature of matter, today constitute a fertile playground for studying a diverse range of physical phenomena, captivating scientific curiosity and attracting substantial industrial interest. In recent years, significant advances have been made in the development of both physically-inspired empirical and machine learning (ML) potentials for computer simulations of atomistic models. However, atomistic simulations of colloidal systems are severely limited by the length- and time-scales achievable with present-day computers. Therefore, computational studies have heavily relied on the use of coarse-grained (CG) models based on effective CG interactions.

In this talk, we discuss recently developed multiscale ML-based approaches for constructing accurate and computationally-efficient CG many-body interaction potentials for complex colloidal systems. We will explore examples where either the CG forces or many-body effective interactions, as extracted from reference fine-grained simulations, are represented by ML models that leverage descriptors of local particle environments. Furthermore, we will discuss how these coarse-graining frameworks may enable the characterization, understanding, and prediction of the structure and phase behavior of relevant colloidal systems through direct and efficient simulations.

Biography: Gerardo is a postdoctoral researcher in Computational Soft Matter Physics at the Debye Institute for Nanomaterials Science, Utrecht University, in the Soft Condensed Matter Group of Prof. Marjolein Dijkstra. He graduated in Chemical Engineering in 2016 from the University of Guanajuato, with a thesis on the development of a statistical-mechanical theory of complex fluid adsorption, under the supervision of Prof. Alejandro Gil-Villegas. From 2016 to 2020, he carried out his PhD studies in computer simulation of selfassembling soft materials at the Multiscale Modelling Group of The University of Manchester, working with Prof. Flor Siperstein and Dr.  Alessandro Patti. After obtaining a PhD in Chemical Engineering and Analytical Science, he joined the Soft Condensed Matter Group at Utrecht University in August 2020. In his research, Gerardo focuses on the development and application of computational methods to the study of complex fluids. His current research encompasses the investigation of the self-assembly of anisotropic particles in bulk and under confinement, and the development of reduced order models for colloidal systems using atomistic simulations and machine learning.

Alexandru Telea

A. Telea portrait
A. Telea

Friday 10:00 - 11:00

Data Visualization: Seeing is learning in high dimensions

In this lecture: How I aim to provide an overview of how to visually explore high-dimensional data actionably and effectively by state-of-the-art methods, avoid errors or misconceptions when using such techniques, so domain experts (like chemists) can make the best use of recent advances in information visualization.

Multidimensional projections (MPs) are one of the techniques of choice for visually exploring large high-dimensional data, such as emerging from measurements or simulations in chemistry, physics, or similar fields. In parallel, machine learning (ML) and in particular deep learning applications are one of the most prominent generators of large, high-dimensional, and complex datasets which need visual exploration. In this talk, I will explore the connections, challenges, and potential synergies between these two fields. These involve “seeing to learn”, or how to deploy MP techniques to open the black box of ML models, and “learning to see”, or how to use ML to create better MP techniques for visualizing high-dimensional data. Specific questions I will cover include selecting suitable MP methods from the wide arena of such available techniques; using ML to create faster and simpler to use MP methods; assessing projections from the novel perspectives of stability and ability to handle time-dependent data; extending the projection metaphor to create dense representations of classifiers; and using projections not only to explain, but also to improve, ML models.

Biography: Alexandru Telea is Professor of Visual Data Analytics at the Department of Information and Computing Sciences, Utrecht University. He holds a PhD from Eindhoven University and has been active in the visualization field for over 25 years. He has been the program co-chair, general chair, or steering committee member of several conferences and workshops in visualization, including EuroVis, VISSOFT, SoftVis, and EGPGV. His main research interests cover unifying information visualization and scientific visualization, high-dimensional visualization, and visual analytics for machine learning. He has authored over 350 papers. He is the author of the textbook “Data Visualization: Principles and Practice” (CRC Press, 2014), a worldwide reference in teaching data visualization.