Tried it at lunch, no one got it "right"

Almost ;-)

“GPT-5 is the first time that it feels like talking to an expert in any topic — a Ph.D.-level expert,” OpenAI CEO Sam Altman said.

Teilchenphysik zum Anfassen im Welios Science Center Von einem Protonen-Fußballspiel bis zu Quarks, die man selbst erzeugen kann, gibt es zahlreiche Möglichkeiten, in einer interaktiven Sonderausstellung im Welios Science Center Teilchenphysik spieleris...

Die Sommerferien gehen in die zweite Hälfte und schön langsam kommt vielleicht auch bei euch mal Langeweile auf. 🙃 Da haben wir einen Tipp: Im @welios.bsky.social gibt es aktuell eine interaktive Ausstellung für Kinder zum Thema Teilchenphysik. www.oeaw.ac.at/news/teilche...

2024 Survey Results

Source:
www.postdocnet.mpg.de/214512/2024-...

...less time for work, contributing to more stress and anxiety. Since those restrictions have been lifted, parents may feel less stressed and have more time to spend at work, alleviating some of the professional stress that comes with being a parent and properly allocating time.

... What this means: This result is quite striking and shows the possible effect that the COVID-19 pandemic had on these results. While we do not know definitively why this is the case, we surmise that during the pandemic, parents had increased caring responsibilities and...

Interesting results from the MPG postdoc survey:

"25% of MPG postdocs are parents. Postdocs who are parents are less likely to be anxious and depressed than postdocs who are not parents, and they are more likely to be satisfied with their lives. This is the exact opposite of our 2022 survey.
...

Simulation-based (or likelihood-free) inference.
There have been quite a few papers recently...

If you liked the recent CMS anomaly detection search, check out this paper for more details & validations!

I particularly like this 'rediscovery' of boosted top quarks with anomaly detection!
Left = before AD, right = after AD

Now at #EPSHEP, @hephy.bsky.social 's Sahana Narasimha talking about "Statistically Learning New Physics from LHC Data"

🎉 Great news: Our Machine Learning and Physical Sciences workshop will be back again this year! 🎉
Keep an eye out for updates on deadlines etc, we will be updating the website soon
ml4physicalsciences.github.io
#ML4PS2025

Now at #EPSHEP, @hephy.bsky.social 's Daohan Wang talking about "Higgs Signal Strength Estimation with a Dual-Branch GNN under Systematic Uncertainties"

Now at #EPSHEP, @hephy.bsky.social 's Robert Schöfbeck talking about "Unbinned machine-learned measurements for the LHC with systematic uncertainties"

Based on arxiv.org/abs/2505.05544

Now at #EPSHEP, @hephy.bsky.social 's Daohan Wang talking about "BitHEP - The Limits of Low-Precision ML in HEP"

Based on arxiv.org/abs/2504.03387

And #HEPML takes the center stage!
Thea Aarrestad receives the #EPS Young Experimental Physics Prize at #EPSHEP

Congratulations!

Greetings from #EPSHEP in #Marseille! Looking forward to great physics talks!

new #COMETA video, supported by the Comprehensive Multiboson Experiment-Theory Action (CA22130)Recording of the COMETA Colloquium on the 25th November 2024 by Tehar Aarrestad. Title: Detecting the rare and complex: Machine Learning for precision and efficiency at the LHC Abstract: As datasets in particle physics get progressively larger, algorithms to swiftly and accurately process this data have become increasingly complex. Machine Learning (ML) has emerged as a solution to tackle several of the challenges experiments face: to efficiently select and reconstruct interesting observational data, to enhance sensitivity to increasingly rare processes and to efficiently generate accurate simulations of complex physical systems. The forthcoming High Luminosity LHC (HL-LHC) will collect roughly ten times the data of the LHC, providing new opportunities for precision measurements of rare processes such as the Higgs self coupling, vector boson scattering and vector boson fusion. In this talk, I will present recent developments in ML techniques applied to event selection, reconstruction, and analysis, and discuss their impact on enhancing searches for these rare processes. I will review the pivotal role of ML in particle physics, from its contributions to the discovery of the Higgs boson in 2012 to its application in real-time event filtering in HL-LHC, where data rates will approach 5% of global internet traffic. Thea Aarrestad - Detecting the rare and complex - COMETA Colloquium 25/11/2024

🚨 API test! 🚨
☄️ #COMETA throwback friday! ☄️
Thea Aarrestad - Detecting the rare and complex - COMETA Colloquium 25/11/2024
☄️☄️ https://www.youtube.com/watch?v=I7y-VCkEQH4 ☄️☄️

ÖAW: 25 neue Stipendien für Forschende aus den USA FZÖ finanziert Akademie-Programm mit 10 Millionen Euro, um Post-Docs unbürokratisch aus den USA zu holen ++ Faßmann: „Richtiges Zeichen in diesen bewegten Zeiten“

25 neue #Stipendien für Forschende aus den #USA 🎓 Mit 10 Millionen Euro finanziert der Fonds Zukunft Österreich das von der ÖAW initiierte neue Programm APART-USA, das vierjährige Post-Doc-Stellen an österreichischen Universitäten oder Forschungseinrichtungen schafft ⤵️
www.oeaw.ac.at/news/oeaw-25...

Delighted to see our work featured in the news! Two key players mentioned but unfortunately not explicitly named: discussions with Kyle Cranmer and Johann Brehmer were crucial in the initial stages. Also Jay, Rafael were central in making it a reality in ATLAS.

Congratulations, that's awesome!

On Wednesday, Team #HEPHY celebrated winning the #HiggsMLUncertaintyChallenge with a great dinner and a little party!
Congratulations again to the entire team!

Was wondering the same thing...
Is P5' still off?

Ich hatte einiges Glück dabei, habs jetzt mit Familie auf eine tenure track Stelle geschafft

Breaking: The muon discrepancy seems to be officially gone. The Muon g–2 experiment at @fermilab.bsky.social
has just announced its final magnetic moment of the muon, improving precision by 1.8x. This is now in agreement with updated calculations published last week by the Theory Initiative ...🧵

Lorentz Local Canonicalization: How to Make Any Network Lorentz-Equivariant Lorentz-equivariant neural networks are becoming the leading architectures for high-energy physics. Current implementations rely on specialized layers, limiting architectural choices. We introduce Lor...

Lorentz Local Canonicalization (LLoCa) is a drop-in replacement that makes any network Lorentz-equivariant. Check out how we apply it to high-energy physics tasks in arxiv.org/abs/2505.20280.

w/ Luigi Favaro, Peter Lippmann, Sebastian Pitz, Gerrit Gerhartz, Tilman Plehn, and Fred A. Hamprecht
1/6

emacs of course!

And a second shoutout to the ground staff at #BRU #BrusselsAirport, who managed to get my luggage over in the same ultra short time frame! Thank you very much as well!

Shoutout to #BrusselsAirlines and especially the crew of #SN2722 today! My connection was tight and an additional delay made it worse. They let me sit in first row for the landing and had detailed instructions where to run to to catch the next flight. Thank you very much!

Our CMS group is searching, if that's of interest to you:

inspirehep.net/jobs/2911525

inspirehep.net/jobs/2911526