Carsten T. Lüth

When: November 5, 18:00 (Berlin Time) Where: Zoom 🙏 3/3

Test-Time Training unlocks powerful new applications by letting the agent adapt and search for the most relevant information to solve the task at hand. Please register prior to this event. Info:  heidelberg.ai/2025/11/05/jon… Sign up: meetup.com/heidelberg-art… 👇Details Below 2/3

💥 Event alert: Test-Time Training Agents to Solve Challenging Problems We’re thrilled to welcome Jonas Huebotter, PhD student at ETH Zurich, to our joint heidelberg.ai / NCT Data Science Seminar series on November 5th at 6 pm (CET). 👇Details Below 1/3

@giffmana @y_m_asano I want to check back in. Would you still be up for holding a talk at heidelberg.ai? It would also be possible to do so via Zoom.

@CarTLueth @y_m_asano Best to send a new one in early autumn. I'm sure i will forget now, and not taking any new talks over the summer.

It was fun with you all in Heidelberg! :)

🚀 Deploying General AI in the Private World What’s holding AI back in the real world? Join Prof. Seong Joon Oh (@ Tübingen) at the heidelberg.ai / NCT Seminar for insights & new research. 📅 Aug 14, 17:00 (CET) 💻 Online — link in comments 👇 #AI #MachineLearning

📣 I'm excited to share that our paper "SURE-VQA: Systematic Understanding of Robustness Evaluation in Medical VQA Tasks" has been accepted to #TMLR ! 📃 Read our paper: openreview.net/pdf?id=qjNdGpg… 💻 Check out the code: github.com/IML-DKFZ/sure-… More ⬇️

If you ever wondered how Augmented Reality is used in games like Pokémon Go, do not miss this event! Can’t make it live? We’ll upload the recording to our YouTube channel afterward. 📺 YouTube: @HeidelbergAi" target="_blank" rel="nofollow noopener">youtube.com/@HeidelbergAi 3/3

He will touch on topics like MapFreeReloc and ACEZero, i.e. how to derive scene representations from very few (one!) or very many images. ℹ️Info: heidelberg.ai/2025/07/16/eri… ✏️ Sign up: meetup.com/heidelberg-art… 👇Details Below 2/3

We’re thrilled to welcome Eric Brachmann (@eric_brachmann), Senior Staff Scientist at Niantic Spatial, Inc. and a leading expert in visual relocalisation & pose estimation, to our heidelberg.ai / NCT Data Science Seminar on July 16th at 4 PM. 👇Details Below 1/3

I am 100% in favor of this. We need to establish a practice where we accumulate our knowledge and actively question it actively, with the individuals involved gaining well-accredited credentials.

We’re thrilled to welcome Eric Brachmann (@eric_brachmann), Senior Staff Scientist at Niantic Spatial, Inc. and a leading expert in visual relocalisation & pose estimation, to our heidelberg.ai / NCT Data Science Seminar on July 16th at 4 PM. 👇Details Below 1/3

In Nashville for my last PhD conference 🥲. Come join today 10:30-12:30 in Hall D (#391) to talk insights, tips and tricks to modify pretraining for representation reuse - scalably. 🚀Joint work w/ @zeynepakata, @dimadamen, @ibalazevic & @olivierhenaff while at @GoogleDeepMind.

There are still posts about 'new papers showing AI models cannot reason'. There are unfortunately problems into how these evaluations were done and also many of those limitations are known, peer-reviewed and published. Here is a simplified version of what's going on as far as I understand: Say you interview me, and you ask me to write all the integers from 1 to 2^n, for n=13. I will tell you, here is a Python program that writes them. (Even if you force a human to write them down, they will almost certainly make a typo somewhere and that is what's happening in LMs due to sampling). If you then score me zero despite my correct Python solution, this is not evidence I am not reasoning, it just means the evaluation is flawed. The fact that transformers cannot generalize in unseen domains is known e.g. n-digit multiplication, maze solving and many other puzzles. It was shown also how to make transformers generalize with recursive self-improvement. Further science on different reasoning puzzles is important but should not be taken out of its narrow scientific context, and should be peer-reviewed not reduced to viral soundbites. On a higher level, when AI researchers say a model is reasoning, it means it scores well on reasoning benchmarks like AIME. We should not map how we think humans reason to how machines reason. What is practically relevant is that if you post-train a reasoning model to do a task, it will do it well, and generalize in unseen questions from that domain. (like getting a Gold medal in a Math olympiad, doing your taxes, giving legal or medical pointers etc.). We do not know of any domain where this does not happen when enough data and RL is thrown on the domain. That means a ton of tasks can be automated. Outside well-defined tasks, the question 'but is it truly reasoning?' is not a meaningful question to me.

@ziv_ravid I would really like to hear your thoughts on this. For me LLMs including Reasoning Models serve as a database which query and access unstructered knowledge, I never expected even the reasoning models to“really“ reason.

My opinion on Apple's "Illusion of Thinking" paper and the pushback against it: It proved exactly what I thought before!

@giffmana @y_m_asano That would be great! I sent you a formal invitation to heidelberg.ai to your "Contact Me" Mail from your website. Subject: heidelberg.ai -- Invitation Or would another way of keeping contact suit you more?

@CarTLueth @y_m_asano would be happy to but at a later point, for now I don't want more travel this summer anymore :)

@willccbb Theoretical physicists are the intellectual embryonic stem cell, I’ve now seen them become ~everything.

This falls in line with: “everything should be as simple as it can be, but not simpler” -- often attributed to Albert Einstein Source: quoteinvestigator.com/2011/05/13/ein…

@giffmana @y_m_asano @giffmana, we would also be very happy to host you for a talk at heidelberg.ai to hear about your latest research. If you would like to revisit Heidelberg again and meet with the research group from which Mint Medical was created as a spin-off, we could do it in person.

@y_m_asano Nice overview slides of your works!

Slides are uploaded here: dropbox.com/scl/fi/g1lqoq6…

This event was jointly organized by: heidelberg.ai & NCT Data Science Seminar series in collaboration with the ELLIS Unit Heidelberg

In case you missed the last heidelberg.ai talky by Prof. Yuki Asano (@y_m_asano) on "Post-Pretraining in Vision, and Language Foundation Models", it is just released on the heidelberg.ai Youtube Channel: youtube.com/watch?v=5UTCjQ…