TGViewer
Channel Public Channel
Rosenkreutz's Chemical Wedding

Rosenkreutz's Chemical Wedding

@chempsycho

I built cool stuff ai4sci and try to share cool things i saw along the way

@RostislavFedorov
Subscribers
402
Photos
111
Videos
7
Links
172
Recent Posts 20 shown
Post #466 92
Over the last couple of month i was working on portin e3nn library to MLX arcitecture, so that you can speed up your faivourite flaivour of MLIP on your local Apple Silicon computer.

On my fairly basic MacBook, there is a 5-6× speedups for forward and backward passes using pure MLX, and up to 10× with custom Metal kernels, compared with e3nn on CPU.

To make it easier to get started with the library, we also ported some of the classic e3nn tutorials: the early tutorials by Tess Smidt, as well as the invariants tutorial by Martin Uhrin and Thomas Hardin.

Library:
https://github.com/lamalab-org/e3nn_mlx

Documentation:
https://lamalab-org.github.io/e3nn_mlx/

Tutorials:
https://github.com/lamalab-org/e3nn_mlx/tree/main/tutorials
  • ❤ 3
Post #465 186
We built a web-app to try out opencaltura directly from the browser! openclatura.org
  • 🔥 3
Post #464 278
a small showcase of how it works internally
  • ❤ 1
Post #463
Channel name was changed to «Rosenkreutz's Chemical Wedding»
Post #462 272
  • 🔥 5
  • 🤯 3
  • ❤ 1
Post #461
Rosenkreutz's Chemical Wedding pinned «I’ve wanted to update this channel for quite a while, but somehow kept postponing it. So, a small update from me: together with Adrian Mirza and Kevin Maik Jablonka at LAMA Lab, we recently released OpenClatura, an open-source tool for deterministic and inspectable…»
Post #460 301
I’ve wanted to update this channel for quite a while, but somehow kept postponing it.

So, a small update from me: together with Adrian Mirza and Kevin Maik Jablonka at LAMA Lab, we recently released OpenClatura, an open-source tool for deterministic and inspectable structure-to-name conversion.

In simple terms, it takes a molecular structure and tries to generate a systematic chemical name while also showing how it arrived at that result.

We’re currently looking for people who can help us test it. Weird molecules, incorrect names, unexpected outputs, OPSIN mismatches, integration ideas, and brutally honest feedback are all very welcome.

You can install it with:

pip install openclatura

GitHub:
https://github.com/lamalab-org/openclatura

Also, I’ll try not to disappear from this channel for quite so long again.
GitHub GitHub - lamalab-org/openclatura Contribute to lamalab-org/openclatura development by creating an account on GitHub.
  • 🔥 6
  • 😱 2
  • 👍 1
Post #459 663

Forwarded from Computational and Quantum Chemistry

🚨 PhD Position in Machine Learning for Orbital-Free Density Functional Theory 🚨

📍 Heidelberg University, Germany

The Institute for Theoretical Physics (ITP) and the Interdisciplinary Center for Scientific Computing (IWR) invite applications for a PhD position in physics, focusing on orbital-free density functional theory (OF-DFT) using machine-learned functionals.

🔬 Research Focus:
Orbital-free density functional theory offers computational accuracy comparable to coupled-cluster methods but at significantly lower computational cost. Currently, precise density functionals are limited, especially for open-shell systems. This project will implement and extend an existing codebase using machine learning techniques to develop high-quality functionals for open-shell systems.

👩‍💻 Candidate Profile:
- Background in Physics, Chemistry, or Computer Science
- Familiarity with condensed matter physics or theoretical chemistry
- Strong programming skills (Python, C, and Torch preferred)
- Motivation for collaborative, interdisciplinary research

👥 Supervisors:
- Prof. Maurits W. Haverkort (ITP)
- Prof. Fred Hamprecht (IWR)

📧 Apply by: April 30, 2025
Send applications to:
👉 m.w.haverkort@thphys.uni-heidelberg.de

Don't miss this exciting research opportunity!

#PhD #MachineLearning #Physics #Chemistry #DensityFunctionalTheory #OFDFT #HeidelbergUniversity #ResearchOpportunity
Post #458 573
Fred Hamprecht is a cool dude
Post #456 896

Forwarded from Computational and Quantum Chemistry

#PhD

Dear colleagues,

The newly formed Semiconductors and Microelectronic systems (SAM) group at TU Berlin has 2 open PhD positions in Neuromorphic Computing and Hardware for Artificial Intelligence (AI). We’re seeking passionate, self-motivated, and creative researchers eager to design and develop novel materials and devices for bio-inspired sensing and computing.

Inspired by the biological brain, our group will work on nanoelectronic materials and devices that host an innovative hardware-based AI—**neuromorphic computing**. This approach aims to overcome the limitations of software-based AI, enhancing energy efficiency, miniaturization, privacy, and scalability.

The SAM group is led by Prof. Priyamvada Jadaun, Chair Professor of Electrical Engineering and Computer Sciences at TU Berlin. Prof. Jadaun also holds a Visiting Scholarship at the University of California, Berkeley, USA, and an Affiliate position at Lawrence Berkeley National Laboratory, USA.

Please forward this opportunity to interested candidates. Application deadline: November 18, 2024.

🔗 More details on positions:
1. https://tub.stellenticket.de/en/offers/188619/
2. https://tub.stellenticket.de/en/offers/188620/

🔗 More on the group:
1. https://pjadaun.com
2. https://www.tu.berlin/en/sam
Post #455 756

Forwarded from Молбиол, биоинформатика, life science

Более 50% ученых покидает науку в течение 10 лет после начала карьеры, из них более 1/3 в течение первых 5 лет после публикации статьи первым автором, причем для женщин эти показатели выше, чем для мужчин. Nature: https://www.nature.com/articles/d41586-024-03222-7

#карьера
Nature Nearly 50% of researchers quit science within a decade, huge study reveals Nature - Twenty years of publishing data across many countries and disciplines show women are more likely than men to leave research.
Post #454 731

Forwarded from Vladimir Shitov

Я тут случайно выяснил, что у нас открыты позиции постдоков в вычислительной биологии, сингл-селл. Позицию пока не рекламили, поэтому есть шанс проконтактировать раньше всех. Особенность в том, что она на 5 лет: это крутая возможность спокойно поделать крутой рисёрч, а заодно и немецкий паспорт получить, если хочется

Очень умные люди вокруг, топовая лаба по методам анализа сингл-селла в мире, натуры и ML-конференции, расположение в большом, но спокойном Мюнхене. Если шарите в ML, интересуйтесь биологией и звучит актуально, подавайтесь: https://jobs.helmholtz-muenchen.de/jobposting/fff4a86d203a3190e66782d373aa4a056478786b0?ref=homepage

#вакансии
Post #453 721

Forwarded from epsilon correct

Вчера суд Массачусетса вынес решение отклонить иск Франчески Джино к Гарварду и коллективу DataColada, который раскрыл фальсификации в нескольких работах Франчески. Этот скандал с фальсификациями – самый громкий за последние несколько лет: Джино была одним из ведущих учёных-бихевиористов, её работы были классикой в области. Дополнительную перчинку придаёт скандалу название её книги, вышедшей за пару лет до разоблачения: "Rebel talent: Why it pays to break the rules at work and in life". Такая вот ирония судьбы. 🤔

Мне кажется решение довольно важным в контексте прецедентности: расследовать чужой фрод в исследованиях – можно и нужно, в науке должно быть больше разоблачений плохих методов и сомнительных практик. Один из моих любимых блогов по теме ведёт Лиор Пахтер, который знатно проезжался по сомнительно известному в узких кругах исследователю графов Альберту-Ласло Барабаши.

А в ваших областях существуют такие правдорубы? Приглашаю обсудить в комментариях. 👀
Post #449 763
Post #447 827
[Verse]
In a world of circuits and screens
We push the boundaries of what's real
Computational chemistry, the ultimate dream
Bringing science to a whole new deal (woah-oh-oh)

[Verse 2]
Robots in lab coats, precision at hand
AI algorithms, guiding our command
Data mining, simulations so grand
Unlocking mysteries, we'll understand (yeah-yeah)

[Chorus]
Welcome to the future, where molecules collide
In virtual realms, where discoveries reside
No beakers, no test tubes, just code and design
In the realm of computation, we redefine
Older posts →

About this channel

How can I read @chempsycho without a Telegram account?
TGViewer shows the public web preview Telegram publishes for Rosenkreutz's Chemical Wedding: recent posts, photos, videos and the subscriber count, with no app, login or account.
How many subscribers does Rosenkreutz's Chemical Wedding have?
Rosenkreutz's Chemical Wedding (@chempsycho) has 402 subscribers on Telegram, refreshed roughly every 30 minutes.
Does Rosenkreutz's Chemical Wedding know I viewed it here?
No. Public channel previews carry no viewer identity, and TGViewer has no accounts or tracking of what you look up.
Threads Profile ViewerView any public Threads profile without an account.Open ThreadLook →Writing with AI? Make it sound human.Metric37 rewrites AI drafts so they read naturally. Free AI detector, 1,500 words free.Try Metric37 →