Cecilia Clementi

Thanks to @freieuniversitaet.bsky.social @tuberlin.bsky.social @einsteinberlin.bsky.social @dfg.de and BMFTR for support

Peering inside the black box by learning the relevance of many-body functions in neural network potentials - Nature Communications Machine-learned force fields are becoming increasingly popular but suffer from their “black-box” nature. Here the authors adapt explainable AI techniques to coarse-grained graph neural network potenti...

Our latest paper on the interpretation on Neural Network Potentials is now published on Nature Communications: nature.com/articles/s41.... Congratulations to Klara Bonneau, Jonas Lederer and all authors. In collaboration with Klaus-Robert Müller's group.

I am still warming up, hopefully some of my pictures will follow soon.

Next step is indeed fine tuning with experimental data! We will be in touch 😀

Thanks to our collaborators: @franknoe.bsky.social @gdefabritiis.bsky.social. Thanks FU Berlin, @einsteinberlin.bsky.social @dfg.de BMBF for support!

Navigating protein landscapes with a machine-learned transferable coarse-grained model - Nature Chemistry The development of a universal protein coarse-grained model has been a long-standing challenge. A coarse-grained model with chemical transferability has now been developed by combining deep-learning m...

Our development of machine-learned transferable coarse-grained models in now on Nat Chem! doi.org/10.1038/s415...
I am so proud of my group for this work! Particularly first authors Nick Charron, Klara Bonneau, Aldo Pasos-Trejo, Andrea Guljas.

Not really fine is for my research group

Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding

Virtual nodes introduced that pass information via attention; reduces oversquashing; improves GNN on molecules; MD22!!

P: arxiv.org/abs/2502.13797

Meet @cecclementi.bsky.social, Einstein Prof. of Physics at @freieuniversitaet.bsky.social 🇩🇪. An ELLIS Fellow, she pioneers #AI4Science & is a member of #ELLISProgram ML for Molecule Discovery. Her research spans Biophysics & Theoretical/Chemical Physics. #WomenInELLIS

@acaruso0.bsky.social

Thanks Simon! I would be happy to chat about it, it may be useful for what your models as well, if you would like to try it.

In collaboration with @franknoe.bsky.social, great work led by Alessandro Caruso and Jacopo Venturin in my group, with precious help from Lorenzo Giambagli and Edoardo Rolando.
Thanks to @einsteinberlin.bsky.social @dfg.de and BMBF for funding!

Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding Graph Neural Networks (GNNs) are routinely used in molecular physics, social sciences, and economics to model many-body interactions in graph-like systems. However, GNNs are inherently local and can s...

Our new preprint is out: Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding.
arxiv.org/abs/2502.13797
We present a new architecture to address the problem of "short-sightedness" of MLFF based on GNN, which outperform in speed and accuracy Ewald-based GNN.

Many thanks to our colloquium speaker Cecilia Clementi @cecclementi.bsky.social from FU Berlin, for providing a deeper understanding of how machine learning and molecular simulations are leveraged to model protein dynamics. Keep an eye out for the video release on the HITS YouTube channel!

Don't miss our next HITS-SIMPLAIX Joint Colloquium on 10 February at 2:00 PM CET! Cecilia Clementi @cecclementi.bsky.social will talk about Modeling Protein Dynamics with Machine Learning and Molecular Simulation. Sign up ow.ly/h0Wl50UQ2kX to join via Zoom or just come by. @simplaix.bsky.social

Or you can come to visit 😀

Thanks!

Thanks !

I made panettone tiramisu with mascarpone in the dough and inserts of pistachio coffee and chocolate. With my loyal sourdough starter.

Awesome paper!
BioEmu from MSR: Scalable emulation of protein equilibrium ensembles with generative deep learning www.biorxiv.org/content/10.1...

Looking forward to hearing more explanainers in tomorrow's reading group session on zoom at 9am PT / 12pm ET / 6pm CET: portal.valencelabs.com/logg

"Being a physicist is a dedication to a quest for this kind of understanding."

John J. Hopfield (2024 Physics Nobel Prize)

"Physics was a point of view: that the world around us with effort, ingenuity, and adequate resources is understandable in a predictive and reasonably quantitative fashion."
🧵
John J. Hopfield (2024 Physics Nobel Prize)

"[…] To me Physics was not defined by subject matter: the atom, the troposphere, a piece of glass, my bicycle, a magnet, these are all incidentally the subject matter."
🧵
John J. Hopfield (2024 Physics Nobel Prize)

Good idea!

I had previously made a series of ceramics piggies 🐷

Thanks! Ideally they need to have some approximate circular symmetry because I would like to throw the main body on the wheel and alter it in a second stage. Maybe an octopus… But right now the real bottleneck is the time for this that I don’t have 🤣

Ahhh! Difficult to throw a snake on the pottery wheel but I could try… maybe a very fat and short snake 🤣

This was a project with @franknoe.bsky.social who helped with the design and shaped the baby whale

In 2024 I made a family of ceramic whales by throwing the main body on the wheel and altering the form. 🐳 Now deciding what other pottery animal I can build in 2025. Open for suggestions 😀
#ceramic #pottery

I tried to make pandoro for the first time. It’s a recipe for pandoro al cappuccino e caramello from an Italian blog I follow. Not fully satisfied with the results but @franknoe.bsky.social is excited to eat it 😀. Next weekend I will try panettone.