We are a computational research group at @OxfordChemistry. We aim to understand the structure of complex inorganic materials on the atomic scale.
Oxford, UK
Joined August 2020
- Tweets87
- Following114
- Followers533
- Likes35
Deringer Group Oxford retweeted
Profs @vl_deringer and @fjduarteg have both been busy hosting a Faraday Discussions meeting in Oxford!
"Data-Driven Chemistry": exploring #ML and #AI for chemical discovery.
You can read about it here: chem.ox.ac.uk/article/farada…
@OxfordChemistry @TrinityOxford @Faraday_D #FD_Data
A key idea behind #FD_Data was to build new connections between the molecular and materials #compchem communities – and so it’s jointly chaired by @fjduarteg & myself. Super excited to now see this idea become reality 😀 @Faraday_D
ALT A photo of Fernanda and Volker in front of the meeting venue
Deringer Group Oxford retweeted
The latest on XPOT, our hyperparameter optimisation tool for ML potentials! New ACE functionality, and insight into validation for automated ML fitting w/ @yux_Zhou and @vl_deringer.
Learn more or have a try:
paper: arxiv.org/abs/2408.00656
xpot: github.com/dft-dutoit/xpot
ALT Text logo of XPOT (python package for hyperparameter optimisation of Machine Learning Potentials)
Deringer Group Oxford retweeted
Hyperparameter optimisation for Atomic Cluster Expansion (ACE) ML potentials – the latest #compchem preprint from @VLDGroupOx! Daniel (@DFT_duToit) and Yuxing (@yux_Zhou) show the advantages of optimisation & importance of careful validation. Read more at arxiv.org/abs/2408.00656
Deringer Group Oxford retweeted
Excited to be attending my first conference as part of @VLDGroupOx at @tyc_london !
Great to be at @tyc_london for the 7th Energy Materials workshop! Here with @VLDGroupOx members Bianca & Natascia 😀 #compchem
ALT A selfie of Volker, Bianca, and Natascia in the lecture theatre
Deringer Group Oxford retweeted
How random is the structure of amorphous silicon? Louise's (@lamrosset's) new preprint addresses this question using fast ML-driven simulations & detailed analysis. Great collaboration as always with @dadrabold! Read more here: arxiv.org/abs/2407.16681
ALT Figure 1 from the preprint linked in the post (see https://arxiv.org/abs/2407.16681 for details).
Deringer Group Oxford retweeted
I’m really excited to be invited to share some recent developments in the “AI in Chemistry” session. It’s great to be returning to Dublin, one of my favourite cities, to connect with so many young researchers to share ideas and results. Look forward to seeing you all there 😁🇮🇪
🌟 Presenting the EYCN Early-Career Program at #ECC9! 🌟
Join us for inspiring career panels, engaging workshops on science communication, and expert networking opportunities. Don't miss out! 🌱🔬
#EYCN #YoungScientists #ECC9 #Conference2024
@EuChemS_Congres @EuChemS
Deringer Group Oxford retweeted
Data as the next challenge in atomistic #MachineLearning - very happy to share this Comment in @NatComputSci 🙂 Thank you Chiheb (@bm_chiheb) & John (@jla_gardner)!
Read more here: nature.com/articles/s43588-0…
ALT A schematic sketch of a dataset (left) and a machine-learning model (right), similar to Fig. 1a of the published Comment linked in the post.
Deringer Group Oxford retweeted
Proud to share some recent work on understanding the atomic scale structure of graphene oxide and how this structure changes during thermal reduction.
Really fun collabaration with colleagues here in Oxford and also an excuse to work with the amazing @m_a_caro and Tigany!
What does graphene oxide (GO) look like on the atomic scale? And how does its structure change during heating?
Excited to share this #compchem preprint – led by Zak (@ZakMachachi), & including a great collaboration with @m_a_caro's group! Read more here: arxiv.org/abs/2405.14814
Deringer Group Oxford retweeted
Delighted to share this #compchem paper from @ZFaureBeaulieu's ongoing DPhil project – a great collaboration with @fausto_martelli at @IBMResearch! Zoé studied #ML models that can classify different forms of amorphous ice. Read more in @JChemPhys doi.org/10.1063/5.0193340
ALT Selected figures from the linked paper, including a schematic (Figure 1) and relevant results (Figures 2b and 4a).
Deringer Group Oxford retweeted
Device-scale atomistic modelling of phase-change memory materials (digital "ones & zeroes") – now published in @NatureElectron! 😀
We showcase #MachineLearning driven #compchem simulations that are relevant to real-world memory devices.
📄🔓 nature.com/articles/s41928-0… (1/5)
ALT Figure 4a-b from the linked paper: a schematic sketch of a memory device, and a simulation of the melting process
Deringer Group Oxford retweeted
What's so special about amorphous calcium carbonate (ACC)? Read about it in our @NatureChemistry paper - a collaboration with @OxColloidGroup & the Goodwin group here at @OxfordChemistry, + US colleagues. Thank you @thomascnicholas & all involved 🎉
nature.com/articles/s41557-0…
Deringer Group Oxford retweeted
Very pleased to present our new pre-print: "Synthetic data for pre-training neural-network potentials" 😌
See 👇🧵 for a quick summary, or have a complete read here: arxiv.org/abs/2307.15714
All thoughts and comments very welcome!
1/
Deringer Group Oxford retweeted
ML acceleration for molecular dynamics in CASTEP! Hot off the press (and "Featured") at @JChemPhys 😀
pubs.aip.org/aip/jcp/article…
#compchem #MachineLearning #OpenAccess - mini thread below (1/5)
ALT An overview of the methodology - Figure 1 of the linked article.
Deringer Group Oxford retweeted
In case you missed it: the fabulous @DFT_duToit's "XPOT" tool for optimising hyperparameters of ML potentials was recently published over @JChemPhys: doi.org/10.1063/5.0155618
Deringer Group Oxford retweeted
#MC16 by @RoySocChem is the best opportunity to network with (potential) collaborators in-person. 💡
In the photo are @JoanaBustamant1 , @NaikAak , Katharina Ueltzen, @MolecularXtal , @vl_deringer, @ZakMachachi, @yux_Zhou and @cer5814012 .
@BAMResearch
@VLDGroupOx
#compchem
Deringer Group Oxford retweeted
New group photo! 😀 Such a pleasure to see team @VLDGroupOx expanding and going strong - thanks to everyone's creative contributions, & to generous support by @UKRI_News @EPSRC and of course @OxfordChemistry
ALT A photo of the Deringer group in June 2023, taken in front of the Chemistry Research Laboratory at Oxford.
Deringer Group Oxford retweeted
A pleasure to welcome Dr Chiheb Ben Mahmoud (@bm_chiheb) as a new postdoctoral researcher @VLDGroupOx!
Chiheb will lead activities on #MachineLearning for spectroscopy & structural characterisation of amorphous materials, as key part of our @UKRI_News Frontier Research project.
ALT A portrait photo of Dr Chiheb Ben Mahmoud on South Parks Road in Oxford.
Deringer Group Oxford retweeted
Very excited that my Master’s project has made it into preprint form!🥳
Read all about it here: arxiv.org/abs/2305.05536
Thank you to my supervisors @vl_deringer, Andrew Goodwin and to my co-authors @thomascnicholas and @jla_gardner for all their hard work and support.
(1/3)
ALT Figure 1 from the linked preprint. Schematic of structural coarse-graining.
Deringer Group Oxford retweeted
A good example of combining machine-learned potentials with chemical bonding analysis (using LOBSTER 🦞!) - both are my favorite tools to study inorganic chemistry 😆!
Special thanks to Stephen and @vl_deringer for fabulous contributions!
Congratulations @yux_Zhou whose work on amorphous red phosphorus is now published in Angewandte Chemie (@angew_chem)! 🥳
We combine ML-driven #compchem & LOBSTER bonding analysis to study this widely-known, yet structurally highly complex material: onlinelibrary.wiley.com/doi/…
ALT Visualisation of the three structural models of amorphous red phosphorus discussed in the paper.
Deringer Group Oxford retweeted
Synthetic data for atomistic #MachineLearning - our first paper in @digital_rsc! 🎉
John (@jla_gardner) and Zoé (@ZFaureBeaulieu) show that a large dataset, itself generated with an ML model, can be useful for supervised & unsupervised learning tasks: pubs.rsc.org/en/content/arti…
ALT The Table of Contents graphic from the linked paper, summarising the approach in a schematic way. See https://doi.org/10.1039/D2DD00137C for details!