Philipp Benner retweeted
Die Investitionen in die Wasserstoffwirtschaft brechen ein. Eine Katastrophe mit Ansage. Davon hängt ab, ob wir Industrieland bleiben können. Gas und Atom sind keine Lösung. Wir haben weder Atom- noch CO2 Endlager. share.google/KzBLUYI7Mzjf2hz…
Philipp Benner retweeted
Matbench Discovery is out in Nature Machine Intelligence @
Paper: rdcu.be/esRFB
Leaderboard: matbench-discovery.materials…
No better time to thank all my co-authors @RhysGoodall, @PhilippBenner2, Yuan Chiang @cyrusyc_tw, @Bowen_D_, Mark Asta, Gerbrand Ceder @cedergroup,
Philipp Benner retweeted
PSA for machine learning force field authors who already have or are planning to submit models to Matbench Discovery.
There's a new dynamic scatter plot on the landing page (matbench-discovery.materials…) that allows comparing all 30 existing models against each other across 2 dozen metrics, hyper-parameters and model design choices.
We hope that as new models and metrics get added to the leaderboard, this will surface empirical insights about which aspects of a model's design space positively or negatively impact certain metrics.
For users of these ML force fields, we hope the dynamic nature of this plot will allow constructing exactly the Pareto front of metrics they care about to make a more informed decision on which potential best suits their research needs.
To make this plot more informative, we encourage future model submission to be as detailed as possible in reporting the hyper-parameters and training procedure chosen for a given model. Existing submissions are welcome to backfill any data that wasn't originally recorded but could be helpful in this plot! There's an existing model schema on how to report e.g. graph construction cutoffs, max neighbor limits, learnings rates, layer counts, number of trainable parameters, etc. which will require extension to capture less common hyper-parameters as well as pre-training and fine-tuning stages in a standardized way. Happy to collaborate with everyone on that!
The next step for this dynamic scatter plot is to report each model's inference time and memory usage at different system sizes. This will become the default values shown on the X axis as they establish a continuously updated cost-accuracy Pareto front for any metric on Matbench Discovery.
Philipp Benner retweeted
Our paper on FIORA is now officially published in @NatureComms!🔓Peer-reviewed and ready to shake up mass spec predictions ⚗️🔨💻📈
Github: github.com/BAMeScience/fiora
Paper: nature.com/articles/s41467-0…
Many thanks to everyone involved 🙌 #MachineLearning #MassSpec #Metabolomics #FIORA
Philipp Benner retweeted
🚀 MLIPX is Live! 🌟
Check out our new open-source code from @BASF for evaluating machine-learned interatomic potentials. Dive into advanced evaluation methods, visualisation tools, and more! Special shout out to @PythonFZ and Sheena agarwal! github.com/basf/mlipx
Philipp Benner retweeted
Join us on our exciting mission to disrupt the molecular sciences with #MachineLearning #AI at
@MSFTResearch AI for Science. Researcher and Engineering positions open, each can be in Berlin DE or Cambridge UK:
aka.ms/ai4science-engineer
aka.ms/ai4science-researcher
Philipp Benner retweeted
EquiformerV2 [1] + DeNS [2] is now the best model on Matbench Discovery [3] (as of Oct. 18, 2024).
Nicely done by Meta FAIR Chemistry team in their work [4].
Equivariance + Transformers + self-supervised learning indeed work pretty well on 3D atomistic data!
[1] arxiv.org/abs/2306.12059
[2] arxiv.org/abs/2403.09549
[3] matbench-discovery.materials…
[4] arxiv.org/abs/2410.12771
Replying to @OpenCatalyst
Our EquiformerV2 (86M) model pre-trained on OMat and fine-tuned on MPtraj and a subset of Alexandria, is state-of-the-art on Matbench discovery across all metrics with an F1 score > 0.9 and an accuracy of 20 meV/atom.
4/x
Philipp Benner retweeted
#Peptide sind eines der Top-#Forschungsthemen in #Medizin, #Pharma und #Analytik. Wir bieten eine Stelle für eine #Promotion in der #Chemie im Bereich Peptidbibliotheken. Vorkenntnisse in #Peptidsynthese sind wünschenswert. Online-Bewerbung: bam.de/umantis/DE/2011.html
@BAMResearch
Philipp Benner retweeted
Good News: An der philologischen Fakultät der @UniLeipzig wurde eine Position für eine:n
Statistikbeauftragte:n geschaffen und ausgeschrieben (TVL13, 50%, unbefristet!).
Hier die Details, falls das was für Euch ist:
uni-leipzig.de/stellenaussch…
Philipp Benner retweeted
Die BAM setzt weltweite Standards für Sicherheit. Bewerben Sie sich jetzt als Teamleitung im Bereich Forschungsdatenmanagement (FDM) (m/w/d)!
➡️ datacareer.de/job/13795/team…
#forschung #stellenangebot #hiring
Philipp Benner retweeted
‼️Submission deadline extended‼️
Abstract registration is now due on Feb 06, 2024 at 23:59 AoE.
Final submission is due on Feb 13, 2024 at 23:59 AoE.
recomb-seq.github.io/papers/
Submit via EasyChair. Select “RECOMBSeq 2024” track within RECOMB conference at easychair.org/my/conference?…
Philipp Benner retweeted
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
Philipp Benner retweeted
🧫💻 Passionate about #Bioinformatics ? Join our @VingronLab for a #PhD project, e.g. in regulatory genomics, single cell data analysis, or interpretation of genomic variation! Apply for our #IMPRS-BAC #gradschool today !
👉 molgen.mpg.de/4571464/vingro… @maxplanckpress
Philipp Benner retweeted
🚀 We're releasing a pretrained universal interatomic potential MACE for high-throughput materials modeling and robust molecular dynamic simulations.
You can access the model for free under the MIT License on #huggingface ! 🤗
huggingface.co/cyrusyc/mace-…
1/4
Philipp Benner retweeted
Matbench Discovery -- An evaluation framework for machine learning crystal stability prediction. (arXiv:2308.14920v1 [cond-mat.mtrl-sci]) #machinelearning #compchem arxiv.org/abs/2308.14920
Philipp Benner retweeted
We have a couple of positions open! 👇
Join us to have some fun with materials and data.
More positions are to be announced in the next few days, one of them being rated to my beloved solar energy materials. ☀️ 💘
Stay tuned!
#FAIRmat is #hiring!
We are looking for
➡️Data Expert for #Biophysics/ Life Science
➡️Materials Synthesis #DataScientist
Read more and apply online: fairmat-nfdi.eu/jobs/jobs-fa…
#BerlinJobs #STEMJobs #MaterialsScience #NFDI
Philipp Benner retweeted
New job: Scientific Software Developer Java at @BAMResearch Berlin. Skills: Software Engineering. Data Science, Data Engineering, Data Analytics. Python, JavaScript. SQL, NoSQL.
datacareer.de/job/11371/wiss…
#Python #SoftwareEngineering #DataScience #dataengineering #DataAnalytics #SQL
Philipp Benner retweeted
Mistle: bringing spectral library predictions to metaproteomics with an efficient search index biorxiv.org/cgi/content/shor… #biorxiv_bioinfo
Philipp Benner retweeted
👉21.2. #BIFOLD Colloquium: Alexandre Tkatchenko @uni_lu @TUBerlin talks about electrons and Machine Learning Force Fields. ℹ️ Information & registration: bifold.berlin/de/bifold-coll…
#ML #MachineLearning #tuberlin #science
Philipp Benner retweeted
We will celebrate the 65th birthday of our director Jürgen Jost with a conference on "Mathematical Concepts in the Sciences & Humanities". Our greatest desire is to share this celebration with a broad audience, so please register and join us online.
🌐mis.mpg.de/calendar/conferen…




























