←
Prof. Amir Karton2,685 posts
Prof. Amir Karton avatar

Prof. Amir Karton

@Lab_initio

iChem Therefore iAm. Founding Director, Institute for Strategic AI (ISA) | Visiting Researcher @MSFTResearch | Professor of Computational & Materials Chemistry

◷ Joined October 2014▧ 244 Photos and videos
Dozent Tobias avatar
⟳ Prof. Amir Karton retweeted
Dozent Tobias@klein_activist·May 5

Finally, a reference dataset large enough to train ML potentials properly. 73k W1-F12 points change what I can model in materials.

Andy_Weeble_Weaver😷⚫🦋andy-weaver.bsky.social🗿 avatar
⟳ Prof. Amir Karton retweeted
Andy_Weeble_Weaver😷⚫🦋andy-weaver.bsky.social🗿@AndrewW66619812·May 4

“The newly-launched Institute for Strategic AI [UNE] has been using the tech to automate the discovery and testing of potential solvents that can isolate components of the silicon wafers efficiently.”

As Australia’s love for #solar threatens to pile into a serious waste stream, researchers are turning to AI to extract more value from retired #panels.. reneweconomy.com.au/aussie-r…

Orbital Industries avatar
⟳ Prof. Amir Karton retweeted
Orbital Industries@Orbital_Ind·Apr 23

We're proud to be part of the beta release of @nvidia's ALCHEMI Toolkit. This toolkit is a unified environment for building custom atomistic simulation workflows, built on the same work that powered speedups to Orb-v3, the latest version of our simulation model Orb. Recap from March’s Orb-v3 update: Up to 1.7x faster inference on large systems (up to 100,000 atoms) and up to 33x faster inference on batched small systems with TorchSim support. Our forthcoming OrbMol-v2 model will leverage components from the toolkit including: → PME electrostatics for periodic Coulomb interactions → The MTK integrator for batched constant-pressure molecular dynamics The result: materially lower inference time.

Post attachmentPost attachment
Stefan Vučković avatar
⟳ Prof. Amir Karton retweeted
Stefan Vučković@stefan_theochem·Apr 23

Very proud of my PhD student @HengZ_921 and joint work with @nagyrpeter and @TCPUniLu: delivering unprecedented DFT accuracy for charged environments (e.g., metalloenzymes). Now in Science Advances: science.org/doi/10.1126/scia… @unifrChemistry

Paola Gori-Giorgi avatar
⟳ Prof. Amir Karton retweeted
Paola Gori-Giorgi@paolagorigiorgi·Apr 22

The latest release of Skala is out! Accuracy largely improved beyond hybrids at semilocal cost! #compchem

Today we're sharing a major Skala update: new paper and model release. Skala is a deep-learned XC functional for DFT: 2.8 kcal/mol on GMTKN55, wins 32/55 subsets & surpasses SOTA hybrids in accuracy at semi-local cost. Paper: arxiv.org/abs/2506.14665 Code: github.com/microsoft/skala

Frank Noe avatar
⟳ Prof. Amir Karton retweeted
Frank Noe@FrankNoeBerlin·Apr 22

Skala is now clearly the best DFT functional in the widely accepted community benchmark GMTKN55, and is faster than other high-accuracy functionals. If you're generating DFT data today, go ahead and use it! @MSFTResearch #AI #Science #MachineLearning

Today we're sharing a major Skala update: new paper and model release. Skala is a deep-learned XC functional for DFT: 2.8 kcal/mol on GMTKN55, wins 32/55 subsets & surpasses SOTA hybrids in accuracy at semi-local cost. Paper: arxiv.org/abs/2506.14665 Code: github.com/microsoft/skala

Mohammed AlQuraishi avatar
⟳ Prof. Amir Karton retweeted
Mohammed AlQuraishi@MoAlQuraishi·Apr 21

Does AlphaFold’s latent space encode only the native state or something like a distribution over conformations? We begin to answer this question with ConforNets, a mechanism for producing diverse states, or very specific ones, via inference-time adaption of OF3p’s latent space👇

We introduce ConforNets, a mechanism for conformational control in AlphaFold3 models - SoTA at producing diverse conformations on every multistate benchmark (N=104) - Novel capability: transfer state from one protein to another Outperforms BioEmu, ConforMix and AFsample3 🧵1/8

Satya Nadella avatar
⟳ Prof. Amir Karton retweeted
Satya Nadella@satyanadella·Apr 23

Thanks to Prime Minister @AlboMP for the warm welcome today. We are making our largest investment in Australia to date, committing A$25 billion to expand AI and cloud capacity, strengthen cybersecurity, and help people and organizations across the country build digital skills.

Post attachment
Chin-Wei Huang avatar
⟳ Prof. Amir Karton retweeted
Chin-Wei Huang@chinwei_h·Apr 22

Exciting new update of Skala. This table should speak for itself: Skala is now the leading density functional in main-group chemistry surpassing previous SOTA hybrid functional at a cheaper, semi-local cost. #compchem #ai4science

Post attachment

Today we're sharing a major Skala update: new paper and model release. Skala is a deep-learned XC functional for DFT: 2.8 kcal/mol on GMTKN55, wins 32/55 subsets & surpasses SOTA hybrids in accuracy at semi-local cost. Paper: arxiv.org/abs/2506.14665 Code: github.com/microsoft/skala

Rianne van den Berg avatar
⟳ Prof. Amir Karton retweeted
Rianne van den Berg@vdbergrianne·Apr 22

Today we're sharing a major Skala update: new paper and model release. Skala is a deep-learned XC functional for DFT: 2.8 kcal/mol on GMTKN55, wins 32/55 subsets & surpasses SOTA hybrids in accuracy at semi-local cost. Paper: arxiv.org/abs/2506.14665 Code: github.com/microsoft/skala

Renana Poranne avatar
⟳ Prof. Amir Karton retweeted
Renana Poranne@Aromaticist·1 Sep 2025

🎶 If you like it - put a ring on it! 🎵 But... where?? 🤔 Alex's (@awahab_pg) new paper maps the rules of the game for pyrene-based systems! 🥳 From Rings to Properties: Understanding the Effect of Annelation on Pyrene pubs.acs.org/doi/10.1021/acs…