iiinova — where your digital self keeps evolving

Tim Duignan @timothyduignan.bsky.social · 20/10/2025

In practice that means use Omol or Omat standards unless you really need to use something different I guess. As those are the biggest data sets out there?

000

Tim Duignan @timothyduignan.bsky.social · 20/10/2025

When generating training data on specific systems the MLIP/NNP community needs to get organized and agree on one particular level of theory, settings and data storage standards as much as possible so we can pool all the data for training foundation/universal models right?

120

Tim Duignan @timothyduignan.bsky.social · 10/10/2025

I have no idea how long this period will last. But it may go faster than we realise. I have no idea what comes after. But I know I really don't want to miss it.

000

Tim Duignan @timothyduignan.bsky.social · 10/10/2025

We may be just at the beginning of a period where individual scientists working with a team of AI agents will be able to get an immense amount done. Particularly for computational tasks that don't require physical experiments.

100

Tim Duignan @timothyduignan.bsky.social · 10/10/2025

When agentic AI started popping up I remember thinking, yeah that is the logical next step but we're not there yet. Its a several years away, well I think its here now and it came much faster then I expected. It's still early days but seems it will have a profound impact.

100

Tim Duignan @timothyduignan.bsky.social · 09/10/2025

We're beginning to understand the properties and formation of this crucially important complex material (the SEI) in great detail using NNPs/MLIPs. Now we can start using that knowledge to really design and engineer it to perfection. Exciting times ahead!

000

Tim Duignan @timothyduignan.bsky.social · 09/10/2025

Then another one from De Angelis et al. showing a very interesting collective ring diffusion process involving six lithium ions in LiF a key component of the SEI. (chemrxiv.org/engage/chemr...)

100

Tim Duignan @timothyduignan.bsky.social · 09/10/2025

I actually observed these pairs in water a while ago with an NNP and worried a lot that they were a hallucination as it is very counter intuitive that two such small cations would pair like that. But looks like it’s real, and really important! (iopscience.iop.org/article/10.1...)

100

Tim Duignan @timothyduignan.bsky.social · 09/10/2025

James Stevenson et al. show (with experimental evidence) that lithium cations can pair up in the battery solvent and that these pairs are what form the SEI. (chemrxiv.org/engage/chemr...)

100

Tim Duignan @timothyduignan.bsky.social · 09/10/2025

The SEI is the thin layer that forms between the surface of the graphite electrode and the liquid electrolyte in a battery and without it lithium ion batteries wouldn’t be possible as the graphite quickly exfoliates. Despite the its importance we know very little about it.

100

Tim Duignan @timothyduignan.bsky.social · 09/10/2025

Couple of very nice new papers on understanding the SEI formation in lithium ion batteries using neural network/machine learning interatomic potentials (NNPs/MLIPs).

120

Tim Duignan @timothyduignan.bsky.social · 18/09/2025

Universal machine learning forcefields beating tailor made classical potentials for zeolites quite convincingly. Great to see all these benchmarking papers! Again demonstrates accurate training data + speed should be key focus now. arxiv.org/abs/2509.07417

061

Tim Duignan @timothyduignan.bsky.social · 09/09/2025

That’s basically the same machine-learning problem NNPs are already solving and theres already many demonstrations they work well for this purpose. Some tricky problems like coupling the coarse-grained and all-atom levels remain, but that seems solvable.

020

Tim Duignan @timothyduignan.bsky.social · 09/09/2025

But yeah at some point you have to use the nano second scale simulations to train coarse grained models integrate out the short range high frequency motions and learn the free energy surface.

100

Tim Duignan @timothyduignan.bsky.social · 09/09/2025

Which means you can brute force carbonic anhydrase and potassium ion channels etc. easily which are at the fast end admittedly but then with some smart enhanced sampling like replica exchange/meta dynamics it should get you the rest of the way to many important discoveries.

110

Tim Duignan @timothyduignan.bsky.social · 09/09/2025

With distilled, optimized NNPs, and new generation of GPU clusters, should be able to approach classical speeds (100s ns/day).

100

Tim Duignan @timothyduignan.bsky.social · 09/09/2025

I think neural network potentials are the eventual pathway to a virtual cell. The accuracy/memory are quite close to where you need them. Timescale is the last real hurdle. But we can port decades of great tools from classical FFs, so it’s becoming more of an engineering problem now.

100

Tim Duignan @timothyduignan.bsky.social · 08/09/2025

arxiv.org/pdf/2508.15614 This is right and it's a big deal. Been waiting my whole career for this point. So many things to simulate!

030

Tim Duignan @timothyduignan.bsky.social · 02/09/2025

Another very interesting benchmarking paper on NNPs. lnkd.in/gWbcTQw8 It seems the models are pretty much there. Very exciting times as these new large datasets continue to be built. Always need more though!

020

Tim Duignan @timothyduignan.bsky.social · 29/08/2025

There’s examples of how to run Md in the examples folder github.com/orbital-mate... to get a solvated protein there download the pdb and use PDB2PQR to add hydrogens and solvent with parmed or find a classical Md paper where they’ve already done this

github.com

GitHub - orbital-materials/orb-models: ORB forcefield models from Orbital Materials

ORB forcefield models from Orbital Materials. Contribute to orbital-materials/orb-models development by creating an account on GitHub.

010

Tim Duignan @timothyduignan.bsky.social · 29/08/2025

Nice recent example this is an important problem for Pharma: chemrxiv.org/engage/chemr... The new models should be much better at this

chemrxiv.org

011

Tim Duignan @timothyduignan.bsky.social · 29/08/2025

MD= molecular dynamics and is basically just a direct simulation of how the molecules behave we haven't been able to do this accurately for any interesting systems accurately enough until now, which we now can because of these models.

111

Tim Duignan @timothyduignan.bsky.social · 29/08/2025

The main limitation is there are many processes that occur on too long a timescales but we can build big molecular dynamics data sets with this model and then train coarse grained models on those to get to the longer time/spatial scales.

110

Tim Duignan @timothyduignan.bsky.social · 29/08/2025

Many people are already doing amazing science with custom built NNPs for many substances. The idea is now they can skip making the training data and building the model and go straight to doing science.

110

Tim Duignan @timothyduignan.bsky.social · 29/08/2025

It kind of has too many applications to list. But generally for any substance you want to know its structural, kinetic and thermodynamic properties. All of them can be derived from MD in principle.

110

Tim Duignan @timothyduignan.bsky.social · 28/08/2025

For generations people have been dreaming of simulating real complex chemistry like enzymes and MOFs starting from nothing but quantum mechanics. Well I really think it's here.

161

Tim Duignan @timothyduignan.bsky.social · 28/08/2025

This means it should be possible to look at crucially important angstrom/picosecond scale phenomena like the hydrogen bond network of water and how it controls reaction dynamics etc. where we really lack adequate tools.

130

Tim Duignan @timothyduignan.bsky.social · 28/08/2025

I'm particularly excited about how close the structure stays to experiment, with no constraints, even though the training data doesn't contain a single protein.

120

Tim Duignan @timothyduignan.bsky.social · 28/08/2025

And you can look at huge systems with them. Like the 20,000 atoms solvated carbonic anhydrase enzyme with many different complex interactions going on. You can do hundreds of thousands of calculations on it with a single GPU in a few days, with no unphysical behaviour.

130

Tim Duignan @timothyduignan.bsky.social · 28/08/2025

It's amazing to me that you can just pick general purpose DFT validation sets and benchmark them like they are a DFT functional and they will normally do a great job out of the box. Often similar to a dispersion corrected GGA or better but orders of magnitude faster.

120

Tim Duignan @timothyduignan.bsky.social · 28/08/2025

Another remarkable jump in accuracy with these new OrbMol models for simulating chemistry. For example, they now quantitatively reproduce the structure of water. But they should be just as applicable for studying a vast range of different liquids.

281

Tim Duignan @timothyduignan.bsky.social · 24/07/2025

Another nice benchmarking paper highlighting the rapid exciting progress of universal MLIPS/NNPs: www.arxiv.org/abs/2507.11806

010

Tim Duignan @timothyduignan.bsky.social · 23/06/2025

Love this combination of LLMs and NNPs, a powerful pair of tools. www.sciencedirect.com/science/arti... Also wonderful to see people picking up Orb so quickly and getting good results!

020

Tim Duignan @timothyduignan.bsky.social · 19/06/2025

This is excellent! arxiv.org/abs/2506.14492

010

Reposted by Tim Duignan

Jan Hermann @hrmnn.net · 18/06/2025

🚀 After two+ years of intense research, we’re thrilled to introduce Skala — a scalable deep learning density functional that hits chemical accuracy on atomization energies and matches hybrid-level accuracy on main group chemistry — all at the cost of semi-local DFT ⚛️🔥🧪🧬

37225

Reposted by Tim Duignan

Rianne van den Berg @riannevdberg.bsky.social · 18/06/2025

So proud of this work with our amazing team 🤩

0143

Tim Duignan @timothyduignan.bsky.social · 18/06/2025

Once we crack the ability to design SEIs rationally batteries will really take off. Such an exciting field! Please share any others I should read?

000

Tim Duignan @timothyduignan.bsky.social · 18/06/2025

4) Finally this one studying interphase formation on lithium metal. (arxiv.org/abs/2506.10944 ) Until now we have known next to nothing about how these incredibly important layers form or even what they are composed of as they show.

arxiv.org

Coupled reaction and diffusion governing interface evolution in solid-state batteries

Understanding and controlling the atomistic-level reactions governing the formation of the solid-electrolyte interphase (SEI) is crucial for the viability of next-generation solid state batteries. How...

100

Tim Duignan @timothyduignan.bsky.social · 18/06/2025

So many nice NNP papers coming out now it is impossible to stay on top of them. Four very cool recent ones:

131

Reposted by Tim Duignan

Joe Greener @jgreener64.bsky.social · 20/03/2025

Interesting work exploring an enzyme reaction with a machine learning potential. Looking forward to much more like this in the next few years.

031

Tim Duignan @timothyduignan.bsky.social · 21/03/2025

Yeah should hopefully be possible

010

Tim Duignan @timothyduignan.bsky.social · 19/03/2025

Remarkably, despite only being trained on small crystals, it manages to spontaneously reproduce the proton relay carrying protons from the zinc-bound water to the His64 residue — a mechanism that took over a decade of careful experimental studies to unravel.

100

Tim Duignan @timothyduignan.bsky.social · 19/03/2025

So Orb has blown me away again. I simulated the carbonic anhydrase enzyme with it: one of the most important and well studied enzymes in biology. (It converts CO2 to bicarbonate and is involved in many diseases and could also be useful for carbon capture.)

230

Tim Duignan @timothyduignan.bsky.social · 25/02/2025

(The trick is to specify brute_force_knn=False in the calculator.) This is more than an order of magnitude larger than some other universal models. I can get over 7,000 steps of molecular dynamics per day with it.

011