
PhD student @Cambridge_Eng, molecular simulation ∩ machine learning | Previously @AstraZeneca and @impchemistry
Cambridge, England
Joined March 2020
- Tweets23
- Following848
- Followers173
- Likes457
Harry Moore retweeted
Interested in automated workflows for alchemical absolute binding free energy calculations? Check out our recent preprint and accompanying Python package, A3FE:
Manuscript: doi.org/10.26434/chemrxiv-20…
Code: github.com/michellab/a3fe
(1/6)
Harry Moore retweeted
Cool results from Moore/Cole/Czanyi showing that NNPs can learn hydration ∆G for organic molecules much more accurately than forcefields (arxiv.org/pdf/2405.18171):
The ability of NNPs to enable accurate physics-based simulation is still underrated @owl_poster @TimothyDuignan
In our latest preprint, we introduce an efficient, alchemical free energy method compatible with maching learned forcefields, enabling calculation of hydration free energies with first principles accuracy! (1/5)
We provide an implementation of our method using OpenMM, check out the preprint! (4/5)
arxiv.org/abs/2405.18171
Thank you to my PI Gabor and Danny @ColeGroupNCL for supervising! (5/5)
1
348
Harry Moore retweeted
Our next Machine Learning for Quantum Matter workshop will be at Imperial College London (White City) on Friday 3rd May.
Registration is now open (£15).
This time a focus on neural network wavefunction ansatz, and machine learning force-fields.
frost-group.github.io/ml4qm/
Harry Moore retweeted
The first time I used this model I was pretty breath taken. The coolest things: with just a few lines, **anyone** can now run decent atomic scale simulations.
Try it out yourself! Details in 🧵
paper: arxiv.org/abs/2401.00096
Our large collaborative effort shows the impressive range of applicability of foundational ML potentials, developed using only open-source data and software. More than 30 applications, including MOFs, catalysis, water, and more simulated with one model.
arxiv.org/abs/2401.00096
Interested in simulating biomolecular systems with machine learned forcefields? Go check out the thread below to see the highlights! ⬇️
Very proud to co-first author this work with @davkovacs10, many thanks everyone involved!
💥I am happy to share our work on next gen transferable ML force fields for bio-organic / molecular chemistry.
MACE-OFF23 is a series of pretrained MACE models accurately describing molecules in vacuum as well as condensed phase with a wide range of tests
arxiv.org/abs/2312.15211
Harry Moore retweeted
I'm biased as a co-author, but this is a must-read on how good ML interatomic potentials can be on materials and applications beyond the limits of their training set.
Open the SI. I dare you. 😉
arxiv.org/abs/2401.00096
Harry Moore retweeted
MACE-OFF23: Transferable Machine Learning Force Fields for Organic Molecules. (arXiv:2312.15211v1 [physics.chem-ph]) #machinelearning #compchem arxiv.org/abs/2312.15211
Harry Moore retweeted
Using the recently added MACE multi-GPU support, @PhilippBenner2 burned 1440 A100 hours to train a 16M param universal MACE on MPtrj that achieves a new SOTA on Matbench Discovery, reaching F1 score of 0.64 in crystal stability prediction! 🧵
Harry Moore retweeted
Excited to present Matrix Function Networks (MFNs), our new non-local GNN architecture that parameterizes many-body non-local effects using the spectrum of learnable graph operators. MFNs' structured non-locality outperforms global attention. arxiv.org/abs/2310.10434
Harry Moore retweeted
PhD funding available for UK Students. Interested in machine learning (ML) for molecular modelling? This PhD project aims to develop machine learning inter-atomic potentials (MLIPs) for molecular materials and organic-inorganic interfaces. Please RT 🤗 jobs.ac.uk/job/DBM614/phd-st…
Harry Moore retweeted
For Cambridge locals, or those that don’t mind the travel for one day:
We are organising a LJC showcase day on October 9th.
It’s a rich research focused event, with several talks, a panel discussion and a poster session (with prizes!).
Sign up: forms.gle/hU8A4Khhno2CiH9w9
Harry Moore retweeted
I am happy to share our new paper showing off all the things we wanted to include in the original MACE manuscript, but didn't have the space for!🧐
We show SOTA on QM9, MD22, ANI dataset and several other MD tasks, like spectra from 50 CCSD(T) calculations
arxiv.org/abs/2305.14247
Harry Moore retweeted
.@VytasGapsys and @jhmchem and colleagues develop an end-to-end relative free energy workflow based on a non-equilibrium switching approach that calculates the binding free energies starting from SMILES strings.
nature.com/articles/s42004-0…
Harry Moore retweeted
📢Please RT. Interested in machine learning for molecular modelling of materials and organic-inorganic interfaces?
Apply for a #PhD supervised by our fantastic new colleague @MagdauIoan at @ChemistryNCL! #compchem #ML
jobs.ac.uk/job/CWO737/phd-st…
Harry Moore retweeted
The first simulation is in from @davkovacs10 @jhmchem and team! Here's a 100ps simulation of a complex of TNKS2 in @openmm_toolkit, with the MACE ML FF for the ligand and MM for the protein/water.
























