Moritz R Schäfer
@MoritzRSchfer1
Followers
178
Following
1K
Media
5
Statuses
198
PhD student at @GroupKaestner. Interested in theoretical chemistry and deep learning. MSc. from @UniPadova and @jlugiessen.
Germany
Joined March 2021
RT @LinusKellner: Apax is a really nice MLIP framework. The GMNN implementation strikes a great balance between speed and accuracy for many….
0
1
0
RT @pfau: tl;dr - IPAM, the Institute of Pure and Applied Mathematics at UCLA, will shut down in a few months if their NSF funding isn't re….
0
109
0
Congrats to co-authors Nico Segreto, @PythonFZ, @GroupKaestner, and Christian Holm!. #machinelearning #jax #compchem #ml4science #moleculardynamics.
0
0
3
Apax is a JAX-based platform built for fast and flexible development of MLIPs. Designed with active learning in mind, it delivers: .- Excellent performance.- Shallow ensemble-based uncertainty estimation.- Enhanced sampling capabilities (e.g., uncertainty-driven dynamics).
2
0
1
New Publication 👏 .We are excited to share our latest paper, "Apax: A Flexible and Performant Framework for the Development of Machine-Learned Interatomic Potentials", published in the JCIM. Read the full paper: Code:
github.com
A flexible and performant framework for training machine learning potentials. - apax-hub/apax
2
3
16
RT @docmilanfar: If you expect people to pay attention and be productive, every meeting room, conference room, classroom, office, and home….
0
14
0
RT @tis930: Last week, we continued @lab_COSMO's study of solid-state electrolytes, focusing on the properties of the LPS surfaces https://….
0
5
0
RT @maxplanckpress: Max Planck President Patrick Cramer @mpgpresident has written an open letter to the President of @Harvard, expressing h….
0
39
0
RT @PythonFZ: Finally got around to updating the ZnTrack docs—took me way too long, but here we are! 🚀 Check them out .
0
1
0
RT @fchollet: It's very common to see a 1.2x-5x speedup over pytorch by simply using a proper compiler (JAX) or custom kernels. But 100x se….
0
5
0
RT @marceldotsci: guess i should mention this: as of this month, i’m honoured to receive a walter-benjamin-fellowship from the DFG to fund….
0
1
0
RT @MrclMllr: Important upgrade for the CEH charge model 📈.We’ve improved accuracy and robustness, and extended it for the actinides to cov….
pubs.acs.org
The Charge Extended Hückel (CEH) model, initially introduced for adaptive atomic orbital (AO) basis set construction (J. Chem. Phys. 2023, 159, 164108), has been significantly revised to enhance...
0
5
0
RT @marceldotsci: new work! we follow up on the topic of testing which physical priors matter in practice. this time, it seems that predict….
0
5
0
RT @lab_COSMO: Join the dark side of the forces (or not!). TL;DR - Using forces that are not conservative is not a great idea. Latest and g….
0
10
0
RT @cgarciae88: JAX things to watch for in 2025 by Grigory Sapunov!. Screenshot thread 🧵 (link at the end). 1. NNX
0
9
0
RT @jigyasa_nigam: Look who came to Boston and brought along our first snow and festivities of the season! So happy to see you again @Miche….
0
2
0
RT @HarvardCCB: Congratulations to @MicheleCeriotti for delivering his E. Bright Wilson Prize Lecture on machine learning for chemistry at….
chemistry.harvard.edu
0
4
0
RT @marceldotsci: New: Fast and flexible range-separated models for atomistic machine learning, spearheaded by Kevin and Philip, with many….
arxiv.org
Most atomistic machine learning (ML) models rely on a locality ansatz, and decompose the energy into a sum of short-ranged, atom-centered contributions. This leads to clear limitations when trying...
0
8
0
RT @SandipDeScience: 🚀 MLIPX is Live! 🌟.Check out our new open-source code from @BASF for evaluating machine-learned interatomic potentials….
0
9
0