Jul 25 2025
In light of the ever growing interest in AI and ML within the context of materials research, I’m guest editing a special issue together with Konstantin Klyukin from Auburn University. More information can be found on the flyer below.
(And yes the robot and diamond are AI generated, though it took some effort to get it to have the right number or arms, hold a diamond and look sideways at the same time. 😉
Permanent link to this article: https://dannyvanpoucke.be/special-issue-revolutions-in-the-integration-of-artificial-intelligence-and-machine-learning-in-carbon-based-materials-research/
Jul 20 2025
With the planned arrival of 2 more PhD students to the group, it is finally time to make work of a group logo. QuATOMs is no longer the small solitary one-man-show, but has grown in the last three years to 5 eager PhDs and already 3 Post-docs have been part of the team.
The green-on-black colorscheme is inspired by the oldschool terminals, while the mage penguin of course combines a reference to Tux as well as the sometimes magical nature of computational materials research.
Permanent link to this article: https://dannyvanpoucke.be/quatoms-group-logo/
Jul 08 2025
For the second year in a row, we are organising a summerschool with the master materiomics program, oriented at students in their second or third bachelor chemistry or physics. Within this summerschool the students are introduced into the various topics which play an important role in materials research. Today, as part of the quantum pillar, I had the pleasure to introduce the students into the world of computational research, with a focus on the application for quantum mechanical modelling. We learned for example, that it is practically impossible to store the wavefunction of a simple small molecule like benzene, it would require more great deal more than a mole of galaxies in mass to store it. With Density Functional Theory on the other hand, you can easily investigate it on a modern day laptop, as you only need the electron density.
Materiomics summerschool of 2025.
Permanent link to this article: https://dannyvanpoucke.be/summerschool-materiomics/
Jun 27 2025
Latex is one of the nicest tools for formatting documents, but also a rabbit hole when you want to get that one small feature to your liking. Many PhD students discover how months quickly vanish when trying to create the perfect template…and the situation does not improve with age :-). In a recent bout of rabbit chasing, I decided that I wanted to be able to have syntax highlighting for python available for a course syllabus I’m planning. Early searches looked hopefull as a simple to use package seems to be available for the job (and even suitable for other programing languages): minted. Which is promoted by the tutorial page on overleaf with regard to syntax highlighting. Installation is simple using the MikTex Console.(Though the latter first destroyed itself during an update, losing access to the Qtframework. But Once Miktex was reinstalled, installing minted is trivial.) Unfortunately, WinEdt had an issue which was sufficiently vaguely defined: it could be not installed, incorrectly installed, missing in the path, not accessible without a shell-escape, missing some environment variable…
After some searching (shell-escape switched on, added to the path, installing additional python support packages –>?seriously?) it became clear that the package was not installed entirely correctly. A .py script isn’t sufficient in windows, and an additional cmd script was needed (as well as python installation, which was a bit amazing as we are talking about latex, so there should be no need for python, but apparently minted is a python library, wrapped up for latex). So long story short, if you want to work with minted, you should add a small cmd script to the binary folder of your latex install (after “installing minted” using MikTex) which calls your python installation to execute the python script latexminted.py.
@echo off "C:\Path\To\Python\Install\python.exe" "C:\Path\To\MikTex\Minted\latexminted.py" %*
Permanent link to this article: https://dannyvanpoucke.be/highlighting-python-in-latex/
Jun 23 2025
MSc Thesis presentation of Brent Motmans and Eleonora Thomas (master materiomics students 2025). Both presenting applications of ML in materials research: Machine Learning particle sizes using small lab-scale datasets (Brent) and development of Machine Learned Interatomic Potentials for the modelling of (the dynamics of) H-based defects in diamond.
Today we had the MSc presentations of the master Materiomics. The culmination of two year of hard study and intens research activities resulting in a final master thesis paper. This year the QuATOMs group hosted two MSc students: Brent Motmans and Eleonora Thomas. Brent Motmans performed his research in a collaboration between the QuATOMs and DESINe groups, and investigated the application of small data machine learning for the prediction of the particle size of Cu nanoparticles. His study shows that even with a dataset of less than 20 samples a reasonable 6 feature model can be created. As in previous research, he found that standard hyperparameter tuning fails, but human intervention can resolve this issue. Eleonora Thomas on the other hand introduced Machine Learned Interatomic Potentials (MLIPs) into the group. By investigating different in literature available MLIPs, she pinpointed strengths and weaknesses of the different models, as well as the technical needs for persuing such research further in our group. As collatoral, she was able to generate a model for H diffusion in diamond, with an MAE for the total energy of <10meV/atom, competing with models like google deep-mind’s GNoME.
While working on their MSc thesis, Brent and Eleonora also applied for fellowship funding for a PhD position, and we are happy to announce both Brent and Eleonora won their grant, and will be starting in the QuATOMs group as new PhD students comming academic year. Eleonora Thomas will be working on the modelling of Lignin solvation, while Brent will work in a collaboration with the HyMAD group on the modeling of hybrid perovskites.
Permanent link to this article: https://dannyvanpoucke.be/msc-materiomics-defences-new-quatoms-members/
Jan 21 2025
One of the most joyful parts of teaching, is when you read a student paper and see their joy of the research shine through.
This year was the second year I taught the course “Machine Learning and Artificial Intelligence in modern materials science“, an elective course in the second master materiomics program. As with my other computational courses, there is a strong hands-on component present in this course: a semester long homework assignment, culminating in a paper and presentation of the work done. The basic idea behind the assignment is simple: Take the QM9 dataset and study it using machine learning and artificial intelligence, incorporating things you learn during the course. In practice this means a lot of coding with for example scikit-learn in combination with using every ounce of physical and chemical intuition they gathered during their previous courses. The absolute freedom generally results in some initial trepidation, but intermediate feedback and the growing understanding that the journey is the the actual goal results in some amazing work.
At the end of the semester, I had three papers before me, which could only be written by these three students (Materiomics is a new program, so having 3 of the 7 students picking a rather hard core computational course is good 😉 ). You could feel their own backgrounds and interests seeping through, as well as the fun they had doing so. There was the engineer who approached the problem from a pipeline perspective, the chemist comparing the efficacy of various fingerprints as features, and the physicist who build a new small fingerprint from scratch creating a linear regression model that outperformed all else having R² =1. The last one is a very nice example of frugal computing, of which we do need more in a world suffering climate change. It was also interesting to see also how three totally different stories also hint at the same underlying properties of the dataset (same target being the hardest to predict), a consistency which provides a level of meta-validation of the results. The students themselves also learned to be critical of their own work by comparing the results of different methods used to attack their own research question.
At the end of this course, it is clear they learned more about artificial intelligence than what is possible by just reading about it. The understood the entire workflow of which training is merely a small part, they learned directly the importance of having good quality data and features, and most importantly they learned that they themselves need to be the I in AI, to be successful…and finally, maybe us four should put our heads together and combine this work into a real research paper, as to celebrate the great research done as a “mere homework-assignment”.
Permanent link to this article: https://dannyvanpoucke.be/machine-learning-and-artificial-intelligence-in-modern-materials-science/
Jan 01 2025
2023 and 2024 have been an intense ride, with both the materiomics program, the tenure track and the research group. Since the previous overview in 2022, the QuATOMs group has seen some growth with the arrival of three new members (Pauline Castenetto, Thijs van Wijk, and Aylin Melan). In addition, Emerick Guillaume defended his PhD on the study of diamond growth.
These are not the only things which happened the last two years, so let us look back at 2023 and 2024 one last time, keeping up with  tradition.
Cover Nature Reviews Physics: Accuracy of DFT
Permanent link to this article: https://dannyvanpoucke.be/review-of-2024/
Sep 22 2024
Mid September in Hasselt means the start of a new academic year. It brings to an end a summer of doing some unrestricted research. Especially the last month has been extremely busy.
Today the first week of academic year ended, and I already had the pleasure teaching quantum mechanics and modelling courses to chemistry and materiomics students. We also welcome 2 MSc materiomics students to the group: Brent Motmans and Eleonora Thomas. Brent will be working on an experimental-theoretical project, where the theoretical side will focus on machine learning of his experimental data. Eleonora on the other hand will be combining DFT and machine learning in her study of diamond. A very warm welcome to the QuATOMs group for both.
Permanent link to this article: https://dannyvanpoucke.be/end-of-summer-and-the-start-of-a-new-academic-year/
Sep 01 2024
During the last year, Esin Aylin Melan worked hard at her MSc Thesis within the QuATOMs group. Her research focus was centered on the impact of strain on the zero-phonon-line of the GeV color center in diamond. This work she presented, together with Thijs van Wijk, at the SBDD conference in Hasselt, and was presented as well at both the BPS and EMRS spring meeting of 2024. Before the summer she gave her (very good and enthusiastic) final presentation of the MSc thesis results, bringing her first real research project to good end. (Paper will follow later 🙂 )
Recently, she also received the great news that she was awarded a bilateral PhD Scholarship between UHasselt & UNamur. So from September first, she has started working on the modeling of color centers in diamond and oxides for the coming four years. Welcome to the QuATOMs team, and congratulations on the scholarship. We look forward to the enthusiasm and insights you’ll bring to the team.
Permanent link to this article: https://dannyvanpoucke.be/new-quatoms-group-member/
Jan 10 2024
| Authors: | Emanuele Bosoni, Louis Beal, Marnik Bercx, Peter Blaha, Stefan Blügel, Jens Bröder, Martin Callsen, Stefaan Cottenier, Augustin Degomme, Vladimir Dikan, Kristjan Eimre, Espen Flage-Larsen, Marco Fornari, Alberto Garcia, Luigi Genovese, Matteo Giantomassi, Sebastiaan P. Huber, Henning Janssen, Georg Kastlunger, Matthias Krack, Georg Kresse, Thomas D. Kühne, Kurt Lejaeghere, Georg K. H. Madsen, Martijn Marsman, Nicola Marzari, Gregor Michalicek, Hossein Mirhosseini, Tiziano M. A. Müller, Guido Petretto, Chris J. Pickard, Samuel Poncé, Gian-Marco Rignanese, Oleg Rubel, Thomas Ruh, Michael Sluydts, Danny E.P. Vanpoucke, Sudarshan Vijay, Michael Wolloch, Daniel Wortmann, Aliaksandr V. Yakutovich, Jusong Yu, Austin Zadoks, Bonan Zhu, and Giovanni Pizzi |
| Journal: | Nature Reviews Physics 6(1), (2024) |
| doi: | web only |
| IF(2021): | 36.273 |
| export: | NA |
| pdf: | <NatRevPhys> |
The cover of this issue shows an artistic representation of the equations of state of the periodic table elements, calculated using two all-electron codes in each of the 10 crystal structure configurations shown on the table. The cover image is based on the Perspective Article How to verify the precision of density-functional-theory implementations via reproducible and universal workflows by E. Bosoni et al., https://doi.org/10.1038/s42254-023-00655-3. (The related paper can be found here.)
Cover Nature Reviews Physics: Accuracy of DFT modeling in solids
Permanent link to this article: https://dannyvanpoucke.be/paper2024_accuracycover-en/