Probabilistic AI School 2026 on August 3-7 in Vilnius 🇱🇹. → https://xitter.cf/t.co/oFrKpOpHPr #ProbabilisticAI #ProbAI
Earth
Joined January 2019
- Tweets361
- Following164
- Followers1.1K
- Likes141
ProbAI = Probabilistic AI ✌️ retweeted
KHIPU 2027
Latin American Meeting in Artificial Intelligence
APPLICATIONS EXTENDED!
September 7, 2026
There are no registration fees, and numerous scholarships for accommodation and transportation are available.
More info at khipu.ai | Please spread the word
SÃO PAULO | 5TH - 9TH | APRIL 2027
Thank you to the Artificial Intelligence Association of Lithuania (@lithuania_ai) for hosting the 8th edition of #ProbAI School in Vilnius, and to Linas Petkevičius (@dxlino) for serving as local chair. 🤗 It was a true pleasure working with you and the local team. 🧵
As a president of @lithuania_ai it was a pleasure that association hosted @probabilisticai summer school in Vilnius, Lithuania 🚀 🚀 🚀 It was a long path which took over a year, which lead to this great event not only for Probabilistic AI School (#ProbAI) participants, but also for Lithuanian ecosystem!
I would like to express gratitude for everyone involved into this large effort, but mainly general co-chairs @tarikssalem, David Baumgartner, Povilas Daniušis, and Mantas Lukoševičius 👏 .
The school was full five day event with presentations, and many activities, which would not be possible without involvement of the speakers: Jonas Arruda, @RutaBinky, @gkdziugaite, Helge Langseth, @saramagliacane, Zita Marinho, @DimitriMeunier1, Thomas Dyhre Nielsen, Silja Renooij, @sereliezer, @vernadec and TA's @t_kusmierczyk, Yanzhe Bekkemoen.
Huge thanks once again to diamond sponsor @vinted and all the partners making this possible: @GoogleDeepMind , Nordic AI Hub, Go Vilnius, @AIML_VU .
Till the next events!
We also echo your thanks to the participants #ProbAI School 2026, lecturers, organizing team, volunteers, sponsors, and partners who made this week possible. Until next time! 🖖
Highlights from Day 5⃣, the final day of #ProbAI School 2026: reproducing kernel Hilbert spaces in ML, bias and safety in AI, and Q&A sessions to wrap up the school. 🧵
Rūta Binkytė (@RutaBinky), Jonas Arruda, Jev Gamper from @vinted and Eliezer (@sereliezer) wrapped up the #ProbAI School 2026 with Q&A session.
Zita Marinho’s RL lecture connects the foundations, MDPs, Bellman/value methods, Q-learning and policy optimization, to model-based RL and probabilistic inference. Remember: good decision-making means reasoning over uncertain futures, not simply maximizing the next reward.
And the day concluded with Mantautas Rimkus’s talk on experimentation and variance reduction at @vinted.
Highlights from Day 3⃣ of #ProbAI School 2026: simulation-based inference, causal machine learning, and a poster session.
Correlation can predict, but decisions require understanding what causes what, using experiments, causal graphs, and discovery methods to distinguish association from effect.
A few highlights from Day 1⃣ of #ProbAI School 2026: three lectures covering probabilistic models, Bayesian workflows, and how deep learning models memorize and generalize.🧵
💻 Eliezer de Souza da Silva (@sereliezer) walked us through Bayesian workflows and probabilistic programming, with code examples showing how mathematical models can be implemented in Pyro and used for prediction.
🧠 Gintare Karolina Dziugaite (@gkdziugaite) closed the day with a keynote on memorization and generalization in deep learning, including how misclassified data can push models toward memorizing individual data points.
