Robo Santa is here! 🎅🤖 We introduce REALM, a real-to-sim validated benchmark for robot manipulation. High-fidelity simulation + aligned control = real-world proxy. Stress-test your policy! 👇 Web martin-sedlacek.com/realm/#takea... GH github.com/martin-sedla... arXiv arxiv.org/abs/2512.19562 1/4

Robo Santa is here! 🎅🤖 We introduce REALM, a real-to-sim validated benchmark for robot manipulation. High-fidelity simulation + aligned control = real-world proxy. Stress-test your policy! 👇 Web martin-sedlacek.com/realm/#takea... GH github.com/martin-sedla... arXiv arxiv.org/abs/2512.19562 1/4

Here we got bookmarks and a CEO publicly mocking users when they ask for something. 🤦‍♂️ I will stick around because I like it so much more than LinkedIn but holy hell my patience is depleting and it needs to improve.

Happy Winter Solstice lovely people!

This head is spinning continuously, but we see it rotating back and forth... ...presumably because of our strong prior expectation that faces are convex. This is a very nice example of the Hollow-Face illusion promoted by Richard Gregory: www.richardgregory.org/experiments/

important

Amazing article criticising LLMs, but not from the classical viewpoint of what they can and what they cannot do. Rather, from the lense of how they intensify/accelerate goal displacement in science. A must read!

🚀 We’re opening 2 fully funded postdoc positions in #NLP! Join the MilaNLP team and contribute to our upcoming research projects. 🔗 More details: milanlproc.github.io/open_positio... ⏰ Deadline: Jan 31, 2026

📢 Want a lighter yet stronger Point Transformer? Meet LitePT ✨ LitePT is a lightweight, high-performance 3D point cloud architecture for a wide range of point cloud processing tasks. 💻Code: github.com/prs-eth/LitePT 🌐Project page: litept.github.io 📰Paper: arxiv.org/abs/2512.13689

🏹 Job alert: 3 PhD positions in AI, Earth Observation, and Science-Policy interface 📍 Vannes 🇫🇷 & Ispra 🇮🇹 📅 Apply by Jan 15th 🔗 https://www-obelix.irisa.fr/job-offers/

We're extremely fortunate to be at a point where it makes economic, and not just environmental, sense to build clean energy. We likely needed this tipping point to actually move off of fossil fuels, as much as I wish we could've accomplished more sooner.

Meta's Perception Encoder Audiovisual (PE-AV), the technical engine that helps drive SAM Audio’s state-of-the-art audio separation. PE-AV integrates audio with visual perception. Its native multimodal support sound detection and richer audio-visual scene understanding.

Hossein Azizpour and I are hiring two research engineers to work with us on #uncertainties in scientific #agents at KTH! Both positions are full-time positions for 6 months. Please apply to both positions if you are interested. Ad 1: www.kth.se/lediga-jobb/... Ad 2: www.kth.se/lediga-jobb/...

Training #robots to pour a drink? 🤖🥛 This project teaches robots to understand human actions and sequences. Using over 1M videos and @eurohpc-ju.bsky.social's @lumi-supercomputer.eu Researchers move robots beyond code to learning by seeing. 👉link.europa.eu/HHn6wR #EuroHPC #AI #Robot

Thank you for the post @eurohpc-ju.bsky.social! 🙏🙏 Check our work at ekazakos.github.io/grounded_vid..., where you can download our code, models, and datasets! Test the video captioning and grounding capabilities of your model on our iGround dataset. 📹🤖

Thank you for the post @eurohpc-ju.bsky.social! 🙏🙏 Check our work at ekazakos.github.io/grounded_vid..., where you can download our code, models, and datasets! Test the video captioning and grounding capabilities of your model on our iGround dataset. 📹🤖

Call for Nominations EgoVis 2024/2025 Distinguished Paper Awards. Published a paper contributing to Ego Vision in 2024/25? Innovative &advancing Ego Vision? Worthy of a prize? DL for nominations 20 Feb 2026 Awards announced @cvprconference.bsky.social #CVPR2026 egovis.github.io/awards/2024_...

🏹 Job alert: Two fully funded PhD positions in Natural Language Processing at University of Leipzig 📍 Leipzig 🇩🇪 📅 Apply by Jan 15th 🔗 https://ellis.eu/research/jobs/2025-12-16-two-fully-funded-phd-positions-in-natural-language-processing

Our ICCV 2025 work, GROVE, made it into a EuroHPC success story! Read the article below. A big thank you to @eurohpc-ju.bsky.social for the compute resources and for writing the article! Also big thanks to the User Support team of @lumi-supercomputer.eu for their generous technical assistance! 🇪🇺🇪🇺🇪🇺

Workshop investigating military, surveillance, dual use, and ethical implications of AI technologies!

Our ICCV 2025 work, GROVE, made it into a EuroHPC success story! Read the article below. A big thank you to @eurohpc-ju.bsky.social for the compute resources and for writing the article! Also big thanks to the User Support team of @lumi-supercomputer.eu for their generous technical assistance! 🇪🇺🇪🇺🇪🇺

Worried about AI’s military uses? We are too. We’re organising an ICLR 2026 workshop on AI research and military applications—dual-use risks, transparency, accountability, and ethical/legal governance & policy. Details + paper submissions: see Noa’s post and visit aiforpeaceworkshop.github.io.

🏹 Job alert: 15 PhD positions in the BRIDGE-AI Doctoral Network (AI-Health) 📍 Various cities in 🇫🇮 🇳🇱 🇧🇪 🇩🇪 🇨🇭 🇦🇹 📅 Apply by Jan 30th 🔗 https://bit.ly/4q6ND4M

This is a very serious initiative. While AGI risk debates get much attention, we should worry more about the immediate danger from AI’s role in automating war and surveillance.

🕹️

𝗘𝘂𝗿𝗼𝗟𝗟𝗠-𝟮𝟮𝗕 launched: fully open 22B multilingual LLM for all EU languages, built by a Europe-wide consortium (ELLIS involved), incl. @andre-t-martins.bsky.social (@ellisunitlisbon.bsky.social) ℹ️ Learn more & access the model: 🔗 https://bit.ly/4pIYVMK 🔗 https://bit.ly/4j33XRI

Building bridges, Not weapons. What does AI really bring to our society? Are we ready to face the elephant in the room? 👉 Join the AI for Peace Workshop @ ICLR 2026!

Today in our TUM AI - Lecture Series we'll have the amazing Ruiqi Gao, Google DeepMind. She'll talk about "𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐠𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐰𝐨𝐫𝐥𝐝 𝐦𝐨𝐝𝐞𝐥𝐬: progress and challenges". Live stream: www.youtube.com/live/CkOSMqw... 7pm GMT+1 / 10am PST (Tue Dec 16th).

👋 Meet the ELLIS Board! This episode features Danica Kragic, Professor at KTH Royal Institute of Technology 🇸🇪. She shares her vision for the future of robotics in the age of LLMs & reasoning and tackles our rapid-fire questions on AI, embodiment, & intelligence. Watch now 👉 youtu.be/-OVIoQnIePU

Great initiative... thanks to the organisers

Details and call for papers are here!

Worried about AI’s military uses? We are too. We’re organising an ICLR 2026 workshop on AI research and military applications—dual-use risks, transparency, accountability, and ethical/legal governance & policy. Details + paper submissions: see Noa’s post and visit aiforpeaceworkshop.github.io.

📢 Announcing the dates for #ICML2026 in Seoul! 🇰🇷 July 6th - 11th, 2026 (Monday through Saturday) 📆 Mark your calendar! ✈️ Book your flights! 😅 ...Start stressing about the deadline!

Amazing article criticising LLMs, but not from the classical viewpoint of what they can and what they cannot do. Rather, from the lense of how they intensify/accelerate goal displacement in science. A must read!

Good morning from a Bohemian fairytale…

Best line ever from Wednesday: Principal: Most people would have FOMO Wednesday: I have FOBI: fear of being included

1 month until deadline! Join us to build your own lab in AI + machine learning research. World-class resources incl. @lumi-supercomputer.eu, generous starting package & professorship affiliation with a university in the world’s happiest country! ➡️ www.ellisinstitute.fi/PI-recruit-2... #hiring

Dolphin-v2 🐬 new document parsing model released by Bytedance huggingface.co/ByteDance/Do... ✨ 3B - MIT license ✨ Works on any document: PDFs, scans, photos ✨ Understands 21 types of content: text, tables, code, formulas, figures & more ✨ Pixel-level precision via absolute coordinate prediction

Thrilled to release 🎯 D4RT (Dynamic 4D Reconstruction and Tracking). 🌟 SOTA results on 4D reconstruction & tracking evals 🚀 Up to 100x faster pose estimation than prior works 📍 A simple, unified interface for tracking, depth, and pose using point-wise decoding 🔗 Learn more: d4rt-paper.github.io

Jina AI's jina-VLM A new 2B vision language model that provides multilingual visual question answering and document understanding among open 2B-scale VLMs.

PaCoRe🔥 a new way to scale test-time compute, released by StepFun huggingface.co/collections/...

Found myself repeating this to several students at NeurIPS: When you’re choosing an internship or a job, what you work on and who you work with matter way more than the logo. Don’t optimize for brands. Become the brand!

Historian Thomas Hughes argued that technologies are malleable when young, then harden. Right now we're still shaping AI, or at least it is being shaped by our institutions, norms & use cases Eventually these systems build a momentum of their own. That is why choices now matter, things are fluid.

TITANS & MIRAS: real continual learning MIRAS = a unifying theory of transformers (attention) and state space models (SSM, e.g. Mamba, RNNs) TITANS = an optimal MIRAS implementation that’s “halfway between” SSM & transformer with a CL memory module let’s dive in! research.google/blog/titans-...

#NeurIPS2026 will be held in Sydney, Australia! #ICML2017 was also in Sydney and was an absolute blast

🏹 Job alert: Explainable AI Methods for Self-Driving Labs at Karlsruhe Institute of Technology 📍 Karlsruhe 🇩🇪 📅 Apply by Dec 31st 🔗 https://bit.ly/49NJsWJ

🏹 Job alert: Postdocs on diversity and inclusion in machine learning competitions at IT University of Copenhagen 📍 Copenhagen 🇩🇰 📅 Apply by Jan 31st 🔗 https://candidate.hr-manager.net/ApplicationInit.aspx?cid=119&ProjectId=181837&DepartmentId=3439&MediaId=5

𝗜𝗻𝘁𝗲𝗿𝗻𝗩𝗶𝗱𝗲𝗼-𝗡𝗲𝘅𝘁: 𝗧𝗼𝘄𝗮𝗿𝗱𝘀 𝗚𝗲𝗻𝗲𝗿𝗮𝗹 𝗩𝗶𝗱𝗲𝗼 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗠𝗼𝗱𝗲𝗹𝘀 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗩𝗶𝗱𝗲𝗼-𝗧𝗲𝘅𝘁 𝗦𝘂𝗽𝗲𝗿𝘃𝗶𝘀𝗶𝗼𝗻 Chenting Wang, Yuhan Zhu, Yicheng Xu ... Limin Wang arxiv.org/abs/2512.01342 Trending on www.scholar-inbox.com

We have publicly shared our "PowerCLIP," a method to align powersets of image sub-region with textual structures for precise image-text recognition. Outperforms several SotA in zero-shot classification, retrieval, robustness, and compositional tasks! arxiv.org/abs/2511.23170