Computation of CCSD(T)-Quality NMR Chemical Shifts via Δ-Machine Learning from DFT NMR spectroscopy undoubtedly plays a central role in determining molecular structures across different chemical disciplines, and the accurate computational prediction of NMR parameters is highly desir...

Take a look at Julius Kleine-Büning's work (from @grimmelab.bsky.social) work on ML-based corrections. IIRC, this massively improves the chemical shifts of 1H spectra. The software should be available online. pubs.acs.org/doi/abs/10.1...

www.emtccm.org/application/

Bad news for Swiss chemists, swissuniversities could not reach a "read and publish" agreement with Wiley journals and we enter a no deal situation. Hopefully the situation get resolved soon!
www.swissuniversities.ch/en/topics/op...

Nobel laureates urge US Senate to reject Robert F Kennedy Jr’s nomination Seventy-seven prize winners sign letter arguing Trump’s DHHS pick will put US public health ‘in jeopardy’

Is this scientists being political? Of course: it has clear political implications. Is it right? To express serious concerns about dangers to public health? No question. So it's OK for scientists to be political? Here that's simply what being a scientist demands.
www.theguardian.com/us-news/2024...

A figure comparing box plot and showing all data points. Group1: normally distributed data. Group2: bimodal data. Group3: Data with 3 modes.

Friends don't let friends make bad graphs repo has been updated!

Friends don't let friends use boxplot for binomial (bimodal) data. Is your box plot hiding something from you?

#DataVisualization

github.com/cxli233/Frie...

The article is fairly nuanced and not really captured by the headline, tho.

Very cool work!

Honestly, its a bit sad that @spotify's wrapped is so dull this year. I used to enjoy it, but now it feels generic and imprecise. No need to replace all humans by models, nothing beats a good music nerd.

Let's talk numbers: To date, since Oct 25th, we have found 395,000 Bluesky posts that link to a research paper or other research object that has a scholarly identifier like a DOI, PubMed ID, URN etc

This is enormously high for a platform that has only been open since Feb to general sign-ups.

A lot of PhD students learned today to keep a version controlled repo of their papers in addition to overleaf.

Is @overleaf down or am I crazy?

This a wonderful opportunity!

Jobs -

Hello! I'm posting this both here and on the X-rated site, let's see where it gets more re-posts 😇. We are looking for a research software engineer to help us develop (even) better code for #compchem #atomicscale #machinelearning. Check out the specs and apply!
www.epfl.ch/labs/cosmo/i...

The frequency of sample sizes in published research is not uniform. Here are fraction of sample sizes across research papers, based on google scholar searches. The most popular sample sizes are 10,50, and 100. 79 is the least popular sample size.

They are smart and will find a way! They will be ok!

Murphy's Law of Molecular Dynamics Simulations: "If there's a rare event you want to sample, all other rare events will be sampled first".

There's a new #RDKit blog post introducing some new functionality that I'm really excited about: doing efficient substructure and similarity searches in very large chemical libraries:
greglandrum.github.io/rdkit-blog/p...
#ChemSky

Overview of definitions for interpretability and explainability for different papers. The gist: the authors have different definitions.

What's the difference between explainability and interpretability? Does the machine learning community have an agreed-upon definition?

No.

There's a great overview, which is from this paper: arxiv.org/abs/2211.08943

My take: I prefer interpretability since the term explainability is too strong.

The community turned out to be quite large, so I'm making a sequel:
go.bsky.app/9H4Ldqx
If you'd like to be in it, or you know someone who should be here, please let me know.

Please follow, share, join, and generally - be supportive! 👩‍🔬

My deep learning course at the University of Geneva is available on-line. 1000+ slides, ~20h of screen-casts. Full of examples in PyTorch.

fleuret.org/dlc/

And my "Little Book of Deep Learning" is available as a phone-formatted pdf (nearing 700k downloads!)

fleuret.org/lbdl/

The strain on scientific publishing: we set out to characterise the remarkable growth of the scientific literature in the last few years, in spite of declining growth in total scientists. What is going on?

direct.mit.edu/qss/article/...

A 🧵 1/n
#AcademicSky #PhDchat #ScientificPublishing #SciPub

Already feels sooo much better, honestly, and it's also very nice to see how most of our community made the move!

To be fair, the starter packs made the move quite easy. I used @marwinsegler.bsky.social and @mbarbatti.bsky.social 's ones. Any other ML+chemistry ones I should check?

Let's see if this platform is actually better than the other site. If there are no sketchy bots and I can see the content of those I follow, then it's already an improvement.