Sign in

Martin Trapp

@trappmartin.eurosky.social
3K followers 677 following 634 posts

Assistant Prof in Prob ML @ KTH 🇸🇪 WASP Fellow & ELLIS Member Ex: Aalto Uni 🇫🇮, TU Graz 🇦🇹, originally 🇩🇪. — trappmartin.github.io — Reliable ML | UQ | Bayesian DL | tractability & PCs

PostsRepliesMedia
Martin Trapp @trappmartin.eurosky.social · 10h
Yeah, some proofs/derivations are just boring and technical without much interesting insights. Others are really insightful, really depends in my experience.
110
Martin Trapp @trappmartin.eurosky.social · 19h
Consider that the term Artificial Intelligence was invented in the US, it’s only fair that it dies there too.
020
Reposted by Martin Trapp
Noa Garcia @noagarciad.bsky.social · 06/10/2026
I am running a study on how computer vision researchers perceive dual-use risks of the field. If you work in CV, and would like to help us, the survey takes less than 3min. Reply or DM me for the link. You will be helping us a lot! Reposts are much appreciated 🙏
31011
Martin Trapp @trappmartin.eurosky.social · 05/10/2026
Yeah, seems like they come in all shapes and colours. I am a little shocked tbh.
010
Martin Trapp @trappmartin.eurosky.social · 05/10/2026
I started reading the papers in my ICLR AC stack, and a surprisingly large amount is complete incomprehensible mambo jumbo with theorem after theorem.
2180
Martin Trapp @trappmartin.eurosky.social · 03/10/2026
When you visit a vegan friend.
010
Martin Trapp @trappmartin.eurosky.social · 03/10/2026
Same here, but as a big Escher fan this looks like a paper I need to read. 😅
020
Martin Trapp @trappmartin.eurosky.social · 03/10/2026
At the end of the day, everyone will be Schmidhubered eventually.
050
Reposted by Martin Trapp
Timnit Gebru @timnitgebru.blacksky.app · 01/10/2026
I'm not sure if y'all saw a CNN report on how the US had almost started WW3 with China based on fabricated information generated by a chatbot, claiming that a Chinese vessel was carrying components of nuclear weapons. That was an "existential threat" to humanity www.theguardian.com/commentisfre...
theguardian.com
Forget ‘superintelligence’: error-prone AI nearly sparked world war three this month | Timnit Gebru and Emily M Bender
The risks of AI aren’t what we think they are, as a recent security incident between China and the United States reveals
10411169
Reposted by Martin Trapp
Pierre Alquier @pierrealquier.bsky.social · 30/09/2026
The proceedings of ICML 2026 are online, including a position paper I took part to: proceedings.mlr.press/v306/papamar... The 29 other authors did all the work 😅 but I'm still proud to share it. Also I "officially" become a co-author of so many people I like a lot, this makes me sincerely happy 😊
proceedings.mlr.press
Position: Agentic AI Orchestration Should Be Bayes-Consistent
LLMs excel at predictive tasks and complex reasoning tasks, but many high-value deployments rely on decisions under uncertainty, for example, which tool to call, which expert to consult, or how man...
3101
Martin Trapp @trappmartin.eurosky.social · 29/09/2026
041
Martin Trapp @trappmartin.eurosky.social · 27/09/2026
I don’t think it’s about being smart (we are all smart), it’s about taking research & research ethics seriously and working hard to do a good job.
020
Reposted by Martin Trapp
Mariya Toneva @mtoneva.bsky.social · 24/09/2026
🚨 Two open postdoc positions in the ERC-funded project BrainAlign at the Max Planck Institute in Saarbruecken, Germany: position #1 on NeuroAI: postdocprogram.mpg.de/node/58480 position #2 on human-centered LLM evaluations: postdocprogram.mpg.de/node/58481 Application deadline: Oct 13, 2026
11911
Reposted by Martin Trapp
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 24/09/2026
Genuinely do not understand how to work in robotics, which has tremendous military applications, without being political. It very obviously involves political choices!
8693
Martin Trapp @trappmartin.eurosky.social · 24/09/2026
Also 50% of my submissions got accepted, so I shouldn’t complain. But still. 😅
120
Martin Trapp @trappmartin.eurosky.social · 24/09/2026
Yes, I totally understand that this happens. But man....
210
Martin Trapp @trappmartin.eurosky.social · 24/09/2026
If you wonder, R2 was also positive (weak accept) but didn't fill out the "justification" section.
010
Martin Trapp @trappmartin.eurosky.social · 24/09/2026
R1: "this is a comprehensive study [...] I recommend its acceptance" R3: "this paper provides a comprehensive analysis of an interesting topic and addresses my concerns in the rebuttal [...]" R4: "My concerns [..] have been resolved." AC: "the reviewers concerns were not sufficiently addressed"
230
Martin Trapp @trappmartin.eurosky.social · 24/09/2026
#NeuRIPs2026
I am once again disagreeing with the AC
160
Reposted by Martin Trapp
David Rügamer @davidruegamer.bsky.social · 24/09/2026
Results for PPT are also available now
002
Reposted by Martin Trapp
David Rügamer @davidruegamer.bsky.social · 24/09/2026
NeurIPS results are out (only Main track for now)
164
Martin Trapp @trappmartin.eurosky.social · 22/09/2026
AISTATS is from now on officially called SISTATS 🐍
031
Reposted by Martin Trapp
Andrew Gordon Wilson @andrewgwils.bsky.social · 21/09/2026
I'm a Program Chair for ICML 2027. We'd love any creative suggestions for a great conference! How to manage the explosion of AI slop and insane submission growth? How to manage reviewing? How to incentivize high quality creative work? How to reduce bureaucracy and overhead?
10257
Reposted by Martin Trapp
Venki Murthy @neurovenki.bsky.social · 15/09/2026
Faculty Job Alert 📣! The Life Sciences Departments in Harvard Faculty of Arts and Sciences seek to recruit a tenure-track professor whose research is at the interface of AI and life sciences, from methodological developments to AI-enabled discovery. Please spread the word!
academicpositions.harvard.edu
Tenure-Track Professor in Life Science and AI
The Life Sciences Departments in Harvard Faculty of Arts and Sciences seek to recruit a tenure-track professor whose research is at the interface of AI and life sciences, from methodological developme...
02122
Reposted by Martin Trapp
Martin Modrák @modrakm.bsky.social · 14/09/2026
I will be looking for a PhD student in Bayesian statistics (in collaboration with Paul Bürkner), located in Prague. Forward to students if you have somebody who could be interested. www.martinmodrak.cz/grammo-call/
martinmodrak.cz
Grammo - Let's Research Bayes!
03324
Reposted by Martin Trapp
Thomas Dietterich @tdietterich.bsky.social · 13/09/2026
We are seeing a new trend in submissions to @arxiv.bsky.social (and presumably to conferences and journals): Authors submitting papers whose contents they likely do not understand. 1/
88624
Reposted by Martin Trapp
Cameron @cameron.stream · 11/09/2026
www.anthropic.com/threat-inte...
anthropic.com
Countering misuse of AI: September 2026 / Anthropic
Case studies from threat actors disrupted between December 2025 and August 2026 across seven areas of harm, from cyber operations to biological misuse.
021
Martin Trapp @trappmartin.eurosky.social · 10/09/2026
LLMs are somehow really good at telling you stuff you already know, but in a different, convoluted way, making it look really smart. Almost like replies by users on Reddit. …wait 🧐
0110
Reposted by Martin Trapp
Cameron @cameron.stream · 09/09/2026
I answered a friend's question about how AI systems might enact control over humans, and why it is important to discuss AI risk. People in the AI industry are becoming more and more scared, so I figured it was worth putting some thoughts down.
cameron.leaflet.pub
AI and Control
Forms of power and control and AI might have over us.
2303
Reposted by Martin Trapp
Conference on Secure and Trustworthy Machine Learning @satml.org · 09/09/2026
The submission server for #SaTML2027 is now open! There are also some changes to the Call for Papers, diffs are indicated on the website. Dates: Mandatory abstract deadline: Sep 22 Paper deadline: Sep 29 satml.org
196
Reposted by Martin Trapp
Andrei Bursuc @abursuc.bsky.social · 06/09/2026
💻 Link: uncertainty-cv.github.io/2026/ 💪 Supported by: @elliot-eu.bsky.social
uncertainty-cv.github.io
UNCV Workshop @ ECCV 2026
ECCV 2026 Workshop on Uncertainty Quantification for Computer Vision.
031
Reposted by Martin Trapp
Andrei Bursuc @abursuc.bsky.social · 06/09/2026
🛟 Reliable & reliability folks @eccv.bsky.social! Join our workshop on Uncertainty Quantification for Computer Vision #eccv2026 We have a super lineup of speakers (from computer vision to MLLMs & LLMs) and cool posters. 🗓️ Date: Tue Sep 8 📌Room: Quality View Hotel - Helsingborg
1137
Martin Trapp @trappmartin.eurosky.social · 27/08/2026
One more reason to use open street maps, where the Gulf of Mexico is … tada … the Gulf of Mexico.
020
Martin Trapp @trappmartin.eurosky.social · 25/08/2026
Excited for the WASP(-HS) workshop on Bias, Representation and Transparency in generative models!
000
Martin Trapp @trappmartin.eurosky.social · 25/08/2026
This is insane!
121
Martin Trapp @trappmartin.eurosky.social · 24/08/2026
Cool stuff, and congrats to them.
000
Reposted by Martin Trapp
The Guardian @theguardian.com · 24/08/2026
World’s oceans hit hottest temperature on record in August
theguardian.com
World’s oceans hit hottest temperature on record in August
Average sea surface temperature outside polar regions was 21.1C on Saturday, latest data finds The world’s oceans were hotter than ever previously recorded on Saturday, according to the latest data. The average surface temperature was 21.1C, with the rising marine heat a stark indicator of the climate crisis. More than 90% of the heat trapped by carbon dioxide from fossil fuel burning enters the oceans. Continue reading...
37766
Reposted by Martin Trapp
LMU Open Science Center @lmu-osc.bsky.social · 21/08/2026
Introducing a Guerilla Approach to Scientific Workflows is Prof. Dr @rmcelreath.bsky.social. Register and join virtually to learn about this fresh perspective on developing scientific workflows and managing research complexity. Register here: bit.ly/4pI2GTC 📆 Wednesday, 9th Sept, 09:00-10:00 (CEST)
Green and white flyer advertising a public lecture titled “A Guerrilla approach to scientific workflows” as part of the OSC Open Science Summer School 2026. Visible on the poster are an image of the male speaker, a QR code, and various logos.
185
Reposted by Martin Trapp
tenuretracker.info @tenuretracker.bsky.social · 21/08/2026
Eberhard-Karls-Universitat Tubingen (@unituebingen.bsky.social) is hiring: Two W3 Professorships in Machine Learning and Intelligent Systems #fullprofessor #ai #machinelearning
132
Reposted by Martin Trapp
Markus Pössel @mpoessel.de · 20/08/2026
Ich versuche gerade ein Gefühl dafür zu bekommen, wie der Fortschritt beim deutschen Netzausbau vorankommt. China hat ja ein ähnliches Problem: elektrische Leistung von Erzeugergebieten in die Industrie- und Ballungszentren zu schaffen. 1/
197
Reposted by Martin Trapp
NeurIPS Europe @neuripseurope.bsky.social · 18/08/2026
We are happy to announce that 28 workshops have been accepted for the Paris event, as part of the 102 accepted NeurIPS workshops: blog.neurips.cc/2026/08/10/a... They will take place on Sat Dec 12 + Sun Dec 13, 2026 (for Paris) The suggested deadline for Workshop submissions is close (Aug 29th)!
blog.neurips.cc
Announcing the NeurIPS 2026 Workshops – NeurIPS Blog
0167
Reposted by Martin Trapp
Ethan Mollick @emollick.bsky.social · 18/08/2026
Some of the earlier work I was referring to: “Across the battery of theory of mind tests, we found that GPT-4 models performed at, or even sometimes above, human levels”
0362
Reposted by Martin Trapp
Roger Watt @rjwatt42.bsky.social · 17/08/2026
So here goes. In academic research there are two highly productive strategies: depth/analysis and breadth/synthesis. Depth is to explore a domain as far down into the detail as you can. Breadth is to explore connections between different domains as wide as you can. 1/n
172
Reposted by Martin Trapp
Clément Canonne @ccanonne.github.io · 17/08/2026
If you think that AI reviews in conferences "to save time" or "be thorough" or "handle the large # of submissions coming due to AI", ask yourself how much it'll cost, and who will pay. If not "gifted" by companies, this is going to be north of $30+/paper. You expect 1000+ submissions? Budget that.
4224
Martin Trapp @trappmartin.eurosky.social · 14/08/2026
I have been using for some years and it does the job for me.
010
Martin Trapp @trappmartin.eurosky.social · 14/08/2026
I have been using PaperLib, which is free and a minimal paper manager, for some time and find it sufficient for most of what I need. I can recommend checking it out. paperlib.app/en/
paperlib.app
Paperlib
An open-source and simple academic paper management tool.
120
Reposted by Martin Trapp
James Dinneen @jamesdinneen.bsky.social · 12/08/2026
Ocean temperatures are hovering hundredths of a degree below the all-time record high with months to go before El Niño reaches full strength. “Basically, everywhere on Earth, we’re having above-normal temperatures in the ocean." New in @nature.com: 🌊🌍🧪 www.nature.com/articles/d41...
nature.com
Ocean temperatures set to break all-time record: what's next?
High sea surface temperatures are expected to weaken ocean carbon sinks and damage coral reefs.
13115
Reposted by Martin Trapp
Aki Vehtari @avehtari.bsky.social · 13/08/2026
All four books I've co-authored are freely available online for non-commercial use: - Bayesian Workflow at avehtari.github.io/Bayesian-Wor... Links to other three books are in the quoted post 👇 (too many books to fit in one post!)
avehtari.github.io
Bayesian Workflow book: Website – Bayesian Workflow book
Website for the Bayesian Workflow book by Gelman, Vehtari, McElreath, et al. — case studies, code, and exercises in R and Stan.
8286134
Martin Trapp @trappmartin.eurosky.social · 12/08/2026
Lea - latex test… $p(y \mid x) \propto p(x,y)$
020
Reposted by Martin Trapp
ESA Space Science @science.esa.int · 12/08/2026
The good news? This isn't goodbye. Two more solar eclipses are heading for Europe: a total one in 2027, an annular one in 2028. That's right, it's #EclipsesOnTourInEurope. Start planning! #EuropeanSolarEclipse 🔭
A map of the Iberian Peninsula, showing where three solar eclipses will cross Spain and Portugal. A purple band marked ‘Total solar eclipse 12 August 2026’ curves from Spain's northwest to central east. An orange band marked ‘Total solar eclipse 2 August 2027’ covers Spain's southern tip and northern Africa. A green band marked ‘Annular solar eclipse 26 January 2028’ crosses the peninsula diagonally from southern Portugal to northeastern Spain.
212640