Research group leader @ Max Planck Institute for Informatics | Interested in ML, Stats, Data Mining, and everything that connects them.
Joined July 2022
- Tweets100
- Following148
- Followers148
- Likes377
Jonas Fischer retweeted
The 2026 Nobel Prize in physiology or medicine goes to the developers of optogenetics. Did you know that optogenetics is based on a discovery made in green algae? Well, now you do! Here’s a comic explainer we did back when it was still new. @PHDcomics #neuroscience @NobelPrize
Jonas Fischer retweeted
Come talk to me about Circuits for Vision Transformers at @eccvconf 10:30, poster number #67
Jonas Fischer retweeted
We are presenting VIGA at ECCV this week. It was the first agentic approach to do this. But now I wonder why we are presenting it given how quickly the field has moved. It is making me reevaluate the role of academic vision research. fugtemypt123.github.io/VIGA-…
Jonas Fischer retweeted
Tomorrow at #ECCV2026, we present our spotlight paper that builds a CLIP-like vision foundation model with an interpretable concept decomposition that is hierarchical and spatially grounded! This work is led by Kai Wittenmayer, with @Amin_Prc , Bernt Schiele, and @JonasFischerML.
Don’t trust your complex AI system? Want interpretable explanations for (almost) any vision task?
🔥check out our language-aligned concept foundation model for vision, a spotlight at ECCV!
🫠 this comes at virtually no performance decrease to the opaque model, and allows for meaningful interventions in LVLMs
Code: github.com/kawi19/CFM
Preprint: arxiv.org/abs/2601.13798
Jonas Fischer retweeted
How Anthropic's new results post would read without the PR:
Claude orchestrated open-source protein design models, PXDesign, RFdiffusion, Genie, BoltzGen, from a 30k-token expert prompt and 12,500 H100-hours of compute, and designed binders against 14 of 15 targets. Hit rates of 22–35% against a 10–15% baseline, where some of those tools already report similar numbers on their own.
The orchestration is genuinely impressive. But the open-source models did most of the lifting, and they came from the Baker lab, Columbia, MIT, ByteDance Seed, and most of them were already wet-lab validated before Claude touched them.
Which also sets the ceiling. All these generators share a single PDB-shaped training distribution, so calling four of them doesn't diversify away the blind spot, since they fail together. The targets that worked are the well-studied ones.
So the valid claim is that an agent can now drive this stack competently in the regime where the stack already works.
Instead, we got this announcement:
Many drugs work by binding to a specific target in the body and blocking or changing what it does. An important first step in the drug development process is designing a molecule that can bind tightly to its target. Traditionally, that's meant weeks or months of expert work per target, sifting through a large number of candidates to identify the few that work.
We wanted to test if Claude could successfully design novel protein binders from scratch (also called de novo design). With a protein design prompt written by a human expert, Claude autonomously designed protein binders against 14 out of 15 targets.
We then worked with Adaptyv Bio and Twist Bioscience, who independently built and tested the proteins Claude designed.
Jonas Fischer retweeted
📷 Deadline extended!
The submission deadline for the WOOP Nectar track at #ECCV2026 has been extended to August 28.
We welcome completed work on world models, evaluation, embodied AI, robotics, and related topics.
CFP: eccv26woop.github.io/#call
Details in the post below
How do we evaluate whether world models are actually useful in downstream systems?
Join us at WOOP—World Models in the Loop—at #ECCV2026
📷 Submission deadline: Aug. 14
📷 eccv26woop.github.io
@olaf_duenkel @NhiPham49 @KenJSimmoteit @thunguyenphuoc @AdamKortylewski
Jonas Fischer retweeted
📢 Deadline extended!
The submission deadline for the WOOP Nectar track at #ECCV2026 has been extended to August 28.
We welcome completed work on world models, evaluation, embodied AI, robotics, and related topics.
Details in the original post below 👇
WOOP @ #ECCV2026 !
As part of the ECCV workshop on World Models in the Loop, we are inviting submissions of papers to our Non-Proceedings track, accessible via eccv26woop.github.io/#call
Organizers: @NhiPham49 @KenJSimmoteit @thunguyenphuoc @AdamKortylewski
See you in Malmö!
Jonas Fischer retweeted
Hiring! Postdoc / Research Leader on human–world interaction, interactive world models and /or human motion & appearance (faces + full body).
virtualhumans.mpi-inf.mpg.de…
Amazing Tübingen ecosystem: @uni_tue @MPI_IS @ELLISforEurope
Feel free to forward to interested people!
Jonas Fischer retweeted
📢 Fully funded PhD @ Helmholtz Munich + TUM
We're hiring PhD students to build trustworthy AI systems for science: explainable AI, VLMs, reliable adaptation, agentic workflows.
⏰ Priority deadline 10 Aug 2026. Application details below 👇
Jonas Fischer retweeted
Have ideas for building and scaling interpretability methods?
Submit to the Explainable Computer Vision Workshop (eXCV) taking place at #ECCV2026! This year, the theme is the role of XAI in the era of foundation models.
@eccvconf
ALT Call for Papers for the Explainable Computer Vision Workshop at ECCV 2026.
Jonas Fischer retweeted
We have 1-2 more extra spots due to new funding -- apply by end of April!
Jonas Fischer retweeted
Can 3D scenes be represented by and rendered from a set of compressed tokens?
It turns out they can and it pairs very well with generative rendering to handle uncertainty!
Make sure to check out @Mohamma68780050's recent work Scenetok, accepted at #CVPR2026.
Links below.
Jonas Fischer retweeted
Our paper “CLIP-Free, Label-Free, Unsupervised Concept Bottleneck Models” has been accepted to #CVPR2026 (Findings). This is one of my strongest PhD works, but at the same time, the most unluckiest. Anyways, Im happy to see its out now! #xai
Paper: arxiv.org/pdf/2503.10981
Jonas Fischer retweeted
We're hiring ~10 research engineers on the Interpretability team at Anthropic. If you're a seasoned ML infra engineer who's excited about understanding what's happening inside frontier models, we'd love to hear from you.
(No prior interp experience needed!)
Jonas Fischer retweeted
Huge thanks to my great co-author @AJesslen, and my three amazing advisors Bernt Schiele, @AdamKortylewski*, and @JonasFischerML*! (*equal advisorship)
📄 Paper: arxiv.org/abs/2503.13429
🌐 Project: phamleyennhi.github.io/cave/
💻 Code: github.com/phamleyennhi/CAVE
#ICLR2026 #XAI #3D
