[{"content":"","date":null,"permalink":"https://digitalbaker.net/tags/announcement/","section":"Tags","summary":"","title":"Announcement"},{"content":"","date":null,"permalink":"https://digitalbaker.net/tags/miai/","section":"Tags","summary":"","title":"MIAI"},{"content":"I am thrilled to announce that I have joined Université Grenoble Alpes (UGA) as a MIAI Research Chair! I will be based at the SIMaP laboratory, focusing on generative models and equivariant neural networks for materials science.\n","date":"1 November 2023","permalink":"https://digitalbaker.net/news/miai-chair/","section":"News \u0026 Updates","summary":"","title":"MIAI Chair appointment at UGA"},{"content":"","date":null,"permalink":"https://digitalbaker.net/tags/","section":"Tags","summary":"","title":"Tags"},{"content":"","date":null,"permalink":"https://digitalbaker.net/tags/uga/","section":"Tags","summary":"","title":"UGA"},{"content":"","date":null,"permalink":"https://digitalbaker.net/categories/","section":"Categories","summary":"","title":"Categories"},{"content":"As part of my work I often bake delicious code, and I always share my recipes. Here are some of the latest.\nmilad Moment Invariants Local Atomic Descriptor\nMilad is a code I created as part of my research to explore ways to encode atomic environments (or any point clouds) in a rotationally invariant fingerprint that can be decoded back into the original environment (modulo global rotation.) If that sounds like something that\u0026rsquo;s up your street then check out the paper, here:\nUhrin, M. (2021). Through the eyes of a descriptor: Constructing complete, invertible, descriptions of atomic environments.\nAiiDA a scalable computational infrastructure for automated reproducible workflows and data provenance\nI was the principal architect of the workflow engine in AiiDA, a highly-scalable and robust workflow engine for scientific workflows. AiiDA has been used in a large number of scientific works with plugins having been written for a large number of community codes, and best of all it\u0026rsquo;s FOSS!\nPapers: Uhrin et al. (2021), Huber et al. (2020).\nmincePy Python object storage with versioning made simple\nI created mincePy because I wanted to collaborate on scientific projects as easily as it is to work together on a Google Doc, or source code in a github repository. MincePy uses MongoDB to let you store, find and keep versions of any Python object (numpy arrays, pandas DataFrames, pytorch networks, etc) in real time from anywhere in the world. pyOS A fresh way to interact with your python objects as though they were files\nPyOS builds on mincePy to provide a bash-like shell where instead of files you have Python objects and instead of a local disk you are connected to a database. kiwiPy Robust, high-volume, message based communication made easy\nKiwiPy is a high-level client for RabbitMQ. It takes the pain out of creating robust message queues which are an essential part of any automated scientific workflow. Uhrin \u0026amp; Huber (2020). plumPy A python workflows library for nested processes with well defined inputs/outputs\nPlumPy is a key component of the AiiDA workflow engine and is used to create Processes with well-defined inputs and outputs and manage their lifecycle in a robust and dependable way. ","date":null,"permalink":"https://digitalbaker.net/codes/","section":"Martin Uhrin","summary":"","title":"Codes"},{"content":"Hi, I’m Martin. #I’m a physicist by training and a computer scientist by heart. I’ve always been fascinated by the \u0026ldquo;hidden\u0026rdquo; rules of the universe—the way a few fundamental symmetries dictate the complex behavior of everything around us. My work lives at the intersection of these two worlds, where I develop the computational tools we need to decode and predict the language of atoms.\nThe Quest for Physical Truth in AI #In recent years, Machine Learning has taken the world by storm, but \u0026ldquo;standard\u0026rdquo; AI often ignores the laws of physics. My passion is in Physics-Informed ML. I don\u0026rsquo;t just want models that find patterns; I want models that understand the physical world.\nI lead the Computational Atomistic Methods and Machine Learning (CAMML) team, where we are particularly obsessed with Equivariant Neural Networks. By baking physical symmetries - like rotations and translations - directly into the architecture of a neural network, we create AI that is not only faster and more data-efficient but fundamentally more \u0026ldquo;truthful\u0026rdquo; to the systems it describes. I’m proud to be an active part of the e3nn community, pushing the boundaries of what’s possible when deep learning respects the laws of nature.\nScience for the Great Transition #Beyond the elegance of the math, I am deeply motivated by the Green Transition. Materials science is the silent engine of the future. Whether it\u0026rsquo;s discovering the next generation of metal-air batteries or designing more efficient catalysts, we are currently limited by how fast we can explore the vast \u0026ldquo;materials space.\u0026rdquo;\nIn the CAMML team, we build Generative Models to solve this. Instead of the slow, traditional process of trial and error, we develop AI that can \u0026ldquo;dream up\u0026rdquo; entirely new atomic structures with the exact properties needed to tackle the global energy crisis.\nLife in Grenoble #Currently, I am based in the beautiful French Alps at the SIMaP laboratory, where I hold a MIAI Chair. Here, surrounded by some of the world\u0026rsquo;s most powerful experimental tools (like the ESRF and ILL), I get to bridge the gap between abstract theory and real-world materials innovation.\nWhen I’m not thinking about equivariant layers or crystal structures, you’ll likely find skiing in the mountains, or sailing on the lakes.\n","date":null,"permalink":"https://digitalbaker.net/","section":"Martin Uhrin","summary":"","title":"Martin Uhrin"},{"content":"Stay informed on our latest lab activities and research breakthroughs.\n","date":null,"permalink":"https://digitalbaker.net/news/","section":"News \u0026 Updates","summary":"","title":"News \u0026 Updates"},{"content":"My papers, preprints and other work, newest first. For citation counts, see my Google Scholar profile.\n2026 Predicting Spin-Crossover Behavior in Metal-Organic Frameworks from Limited and Noisy Data Using Quantile Active Learning Ashna Jose, Emilie Devijver, Martin Uhrin, Noel Jakse, Roberta Poloni Preprint arXiv:2603.03545 arXiv MC3D: The Materials Cloud computational database of experimentally known stoichiometric inorganics Sebastiaan P Huber, Michail Minotakis, Marnik Bercx, Timo Reents, Kristjan Eimre, Nataliya Paulish, Nicolas Hörmann, Martin Uhrin, Nicola Marzari, Giovanni Pizzi Digital Discovery 5(3), 1114–1131 Link 2025 Reproducible container solutions for codes and workflows in materials science Dylan Bissuel, Léo Orveillon, Benjamin Arrondeau, …, Martin Uhrin, … (19 authors) Advanced Engineering Materials, e202503185 Link Predicting first-principles Hubbard parameters with equivariant deep learning Austin Zadoks, Martin Uhrin, Luca Binci, Lorenzo Bastonero, Cristiano Malica, Iurii Timrov, Nicola Marzari Conference talk APS Global Physics Summit 2025 Link Machine learning Hubbard parameters with equivariant neural networks Martin Uhrin, Austin Zadoks, Luca Binci, Nicola Marzari, Iurii Timrov npj Computational Materials 11(1), 19 Link 2022 Toward a unified description of battery data Simon Clark, Francesca L Bleken, Simon Stier, …, Martin Uhrin, … (11 authors) Advanced Energy Materials 12(17), 2102702 Link e3nn: Euclidean Neural Networks Mario Geiger, Tess Smidt, M Alby, …, Martin Uhrin, … (19 authors) Preprint arXiv:2207.09453 arXiv 2021 Data management plans: the importance of data management in the BIG‐MAP project Ivano E Castelli, Daniel J Arismendi‐Arrieta, Arghya Bhowmik, …, Martin Uhrin, … (30 authors) Batteries \u0026amp; Supercaps 4(12), 1803–1812 Link Through the eyes of a descriptor: Constructing complete, invertible descriptions of atomic environments Martin Uhrin Physical Review B 104(14), 144110 Link Common workflows for computing material properties using different quantum engines Sebastiaan P Huber, Emanuele Bosoni, Marnik Bercx, …, Martin Uhrin, … (26 authors) npj Computational Materials 7(1), 136 Link OPTIMADE, an API for exchanging materials data Casper W Andersen, Rickard Armiento, Evgeny Blokhin, …, Martin Uhrin, … (56 authors) Scientific Data 8(1), 217 Link Workflows in AiiDA: Engineering a high-throughput, event-based engine for robust and modular computational workflows Martin Uhrin, Sebastiaan P Huber, Jusong Yu, Nicola Marzari, Giovanni Pizzi Computational Materials Science 187, 110086 Link 2020 Materials Cloud, a platform for open computational science Leopold Talirz, Snehal Kumbhar, Elsa Passaro, …, Martin Uhrin, … (20 authors) Scientific Data 7(1), 299 Link AiiDA 1.0, a scalable computational infrastructure for automated reproducible workflows and data provenance Sebastiaan P Huber, Spyros Zoupanos, Martin Uhrin, … (22 authors) Scientific Data 7(1), 300 Link kiwiPy: Robust, high-volume, messaging for big-data and computational science workflows Martin Uhrin, Sebastiaan P Huber Journal of Open Source Software 5(52), 2351 Link 2019 A High‐Throughput Computational Study Driven by the AiiDA Materials Informatics Framework and the PAULING FILE as Reference Database Martin Uhrin, Giovanni Pizzi, Nicolas Mounet, Nicola Marzari, Pierre Villars Book chapter In: Materials Informatics: Methods, Tools and Applications, pp. 149–170 (Wiley-VCH) Link 2016 Single-layered Hittorf’s phosphorus: a wide-bandgap high mobility 2D material Georg Schusteritsch, Martin Uhrin, Chris J Pickard Nano Letters 16(5), 2975–2980 Link 2015 Understanding the Structure of Materials at the Intersection of Rationalisation, Prediction and Big Data Martin Uhrin PhD thesis PhD thesis, University College London Link 2014 Predicting non-square 2D dice probabilities G AT Pender, M Uhrin European Journal of Physics 35(4), 045028 Link 2011 The MOLDY short-range molecular dynamics package Graeme J Ackland, K DʼMellow, SL Daraszewicz, DJ Hepburn, Martin Uhrin, K Stratford Computer Physics Communications 182(12), 2587–2604 Link ","date":null,"permalink":"https://digitalbaker.net/publications/","section":"Publications","summary":"","title":"Publications"},{"content":"You can download a full CV here.\nResearch Positions #MIAI Research Chair, Université Grenoble Alpes, France November 2023 - present\nPostdoctoral Researcher, EPFL, Switzerland March 2021 - September 2023\nPostdoctoral Researcher, DTU Energy, Danish Technical University October 2019 - March 2021\nPostdoctoral Researcher, EPFL, Switzerland January 2015 - January 2019\nEducation #Ph.D. in Physics, University College London (2015)\nM.Phys in Computational Physics, Edinburgh University (2010)\nIndustry Positions #Software Engineer, Microsoft Game Studios (Rare), UK Spring 2007 - Spring 2008 Worked on the Xbox 360 game \u0026lsquo;Banjo-Kazooie: Nuts \u0026amp; Bolts\u0026rsquo;!\n","date":null,"permalink":"https://digitalbaker.net/cv/","section":"Martin Uhrin","summary":"","title":"Résumé"}]