Robert Haase

Do you enjoy exploring embeddings interactively as much as I do? Check out _embeddoor_! I recently vibe-coded it while watching too many Harry Potter movies 😃
Feedback welcome as more features will certainly come! github.com/haesleinhuep...

Interested in Zebrafish brain imaging? We released a 3D printed mold that's is super helpful for us. Check it out

Customizable FDM-based zebrafish embryo mold for live imaging
www.biorxiv.org/content/10.1...

Work by Marcela and Jaakko in the lab

As part of the #NFDI4DS conference, @haesleinhuepf.bsky.social from @nfdi4bioimage.bsky.social gave a presentation on 'How LLMs impact BioImage Data Science', which is now available at #Zenodo. 🍁

👉 Learn more about the use of LLMs and code generation for bioimage analysis: doi.org/10.5281/zeno...

Today announcement in the train: “Ladies and gentlemen, after our train driver once again drove like a young god, we are now 8 minutes early in Bitterfeld.”
The whole carriage bursts out laughing. Cudos @deutsche-bahn.bsky.social

At the end of September @haesleinhuepf.bsky.social, training coordinator @scadsai.bsky.social held the AI competence training AI4psychology which focused on fundamentals, efficient application, risks, and legal framework of AI models and systems.
Find the slides here:
👉 doi.org/10.5281/zeno...

We just released a new major version of TrackMate (v8), the cell and organelle tracking plugin of Fiji.

It ships many new features, detailed below, but that are articulated around the following:

Collaboration is key to the success of #NFDI from the very beginning. Here is the next video in our strategy series with York Sure-Vetter – this time it's about working together. 🤝

➡️ About the NFDI strategy: www.nfdi.de/verein/
➡️ About the sections: www.nfdi.de/sektionen/

#fdm #rdm #researchdata

Rendering of a full cochlea (left) with three stains (PV, VGlut3, CTBP2) shown in three different colors (red, blue, cyan). The whole cochlea is a spiral shaped structure, with spiral ganglion neurons (SGNs) in the inner helix, labeled by PV and inner hair cells (IHCs) in the outer helix, labeled by Vglur3. The figure also shows zoom ins. The right hand side shows segmentation results for SGNs, IHCs (represented by colored masks) and synapse detections (represented by colored dots).

Preprint alert! CochleaNet, our framework for analyzing light-sheet data of the cochlea. It consists of three networks to segment spiral ganglion neurons, inner hair cells, and to detect synapses. See rendering of a full cochlea in the image, find the preprint at doi.org/10.1101/2025....

Presentation slide: "Working with data and code like a pro – Data Literacy, Open Science and Collaborative Coding and CI/CD," by Mike Berger, Pia Voigt, and Matthias Täschner at the ScaDS.AI Dresden/Leipzig General Assembly in Leipzig, 6 September 2023; with logos for ScaDS.AI Dresden/Leipzig, the Competence Center for Interdisciplinary Data Science, and Technische Universität Dresden, slides are reusable under CC BY 4.0, and funding from the Federal Ministry of Research, Science and the Arts.

Now on #Zenodo: "Data literacy, Open Science and Collaborative Coding and CI/CD" slides from the @Sca_DS GA workshop presented by Mike Berger, Pia Voigt and Matthias Täschner. 
Share, copy, distribute & adapt the work with proper credit to creators!
🔗 https://zenodo.org/records/17130780

Long-context prompting 🚀

Aligning machine and human visual representations across abstraction levels - Nature Aligning foundation models with human judgments enables them to more accurately approximate human behaviour and uncertainty across various levels of visual abstraction, while additionally improving th...

What aspects of human knowledge do vision models like CLIP fail to capture, and how can we improve them? We suggest models miss key global organization; aligning them makes them more robust. Check out LukasMuttenthaler's work, finally out (in Nature!?) www.nature.com/articles/s41... + our blog! 1/3

Join our 34th Living Lab Lecture about "Demystifying Data Stream Processing"!

🗣️ Apurv Deepak Kulkarni (@scadsai.bsky.social / @tudresden.bsky.social) 
📅 November 20 | 11 AM
📍 Online (https://tud.link/i8zf)

👉 https://scads.ai/event/living-lab-lecture-series/demystifying-data-stream-processing/

Slide titled "Explainable Machine Learning" by Robert Haase; ScaDS.AI Dresden/Leipzig with logos of TUD Dresden and Leipzig University; CC‑BY‑4.0 reuse note; date September 24, 2025.

Would you like to learn more about #XAI? 
Take a look at @haesleinhuepf.bsky.social's presentation on "Explainable Machine Learning" on the @scadsai.bsky.social #Zenodo page. It covers the basics, including SHAP analysis and Grad-CAM as examples.

👉 https://f.mtr.cool/mzmitnigaa

Seeing our good old AURA-Net from 2021 (doi.org/10.1109/ISBI...) end up 2nd on a 2025 U-Net variants leaderboard was not on my bingo card for this year, but here we are
(source: doi.org/10.48550/arX...)

JupyterCon!

a black background with green text that says `` hello , world '' ALT: a black background with green text that says `` hello , world ''

This is the eXplainable AI research channel of the machine learning group of Prof. Klaus-Robert Müller at Technische Universität Berlin @tuberlin.bsky.social & BIFOLD @bifold.berlin.
Let's connect!
#XAI #ExplainableAI #MechInterp #MachineLearning #Interpretability

Poster session at globias2025 event.

#GloBIAS2025 is ongoing with amazing speaker lineups at Kobe city

Beautiful location, amazing food and warm bioimage analysis community. What wlse does one need? 😊
Thanks to @globias.bsky.social for this event

Using #LLMs for science👨‍💻💬? We have a fun tool for you: introducing #folder2md4llms. It converts an entire folder of docs/code/pdfs/pptx into a single markdown file that you can upload to your chatbot of choice. We use it to talk with our publications database or code repos
github.com/HenriquesLab...

coral-nlp/german-commons · Datasets at Hugging Face We’re on a journey to advance and democratize artificial intelligence through open source and open science.

We just released "German Commons", the largest openly-licensed German text dataset for LLM training: 154B tokens with clear usage rights for research and commercial use.

huggingface.co/datasets/coral-nlp/german-commons

Episode 17 - perhaps my favorite yet. How do you write a forum.image.sc post that will get you the answers you need without revealing information you don't want to tell? @erinweisbart.bsky.social and I go through what the experts need to know to help you. Post your sci q's today on forum.image.sc !

I just finished my @unileipzig.bsky.social #LunchLecture. Thanks everyone for the vivid discussion and Nicole Koburger for hosting us!
The slides are online #openaccess 🧑‍🏫🔓 zenodo.org/records/1742...

Scientific Developer or Full Stack Developer The European Molecular Biology Laboratory (EMBL) is a world leader in scientific research and at the forefront in the development of cutting-edge technologies. EMBL core facilities and scientists gene...

One more open position at the Data Science Centre here! This one is working on our recently open-sourced ELN/LIMS/all-in-one solution LabID (gitlab.com/lab-integrat...). If you have either full stack or scientific workflows experience, check it out! embl.wd103.myworkdayjobs.com/en-US/EMBL/j...

Presentation slide titled “AI Insights MCP – Model Context Protocol” by Oliver Welz (ScaDS.AI). The slide shows a person at a laptop connected by a plug to a robot, symbolizing human–AI interaction. The ScaDS.AI Dresden/Leipzig Living Lab logo appears in the top right corner. The background is light beige, and the slide is marked with a CC-BY 4.0 license.

Missed our #ScaDSAIMeetup on AI insights into the Model Context Protocol (#MCP)? 
All talks, workshop materials & repos are now on #Zenodo!

🎤 Talks: https://doi.org/10.5281/zenodo.17158643
🛠️ Workshop: https://zenodo.org/records/17158770"

The next @unileipzig.bsky.social's  #LunchLectures will take place on October 23, this time featuring @haesleinhuepf.bsky.social  from @scadsai.bsky.social . He will give a lecture on “How generative AI influences research and teaching.”

👉 https://f.mtr.cool/srahlhpary

As part of our monthly meeting, we do a run-down of all* the new #bioimageanalysis tools that have been published in the last couple of weeks on BioRxiv**, so I thought I'd start sharing here as well.

*I will miss some, apologies if I miss your paper

Together with the @iom-leipzig.bsky.social, @haesleinhuepf.bsky.social & Matthias Täschner hosted the #AISchool2025 last week in Leipzig. The Materials were provided by @scadsai.bsky.social.

👉 https://f.mtr.cool/yzmimoxbvz

#MachineLearning #DataVisualization #HPCWorkflows

Data Management Plan:
* We will keep all data and literature in open browser tabs sychronized between the computers of all project members using the same Google Chrome profile.
* By the end of the project we publish the username and password of that profile under a CC-BY 4.0 license.

"LLMs are here to stay. [...] The best developers won’t be the ones who write every line by hand." 👍

Image Analysis Specialist/Engineer/Coordinator, Helsinki Bioimaging Image Analysis Specialist/Engineer/Coordinator, Helsinki Bioimaging

🚀Join Helsinki BioImaging as an Image Analysis Specialist🖥️to support cross-disciplinary image analysis projects. If you love coding and are interested in working at the crossroads of computer science and biology, apply by 🗓️ November 15, 2025⤵️
jobs.helsinki.fi/job/Helsinki...
#jobopening

#Forschungsbewertung im Fokus

Die Veranstaltung am 10.09.25 bot Raum für Austausch: Warum sind Reformen nötig? Welche Chancen und Herausforderungen ergeben sich daraus? Und welchen Mehrwert bietet eine Mitgliedschaft in der internationalen CoARA – Coalition for Advancing #ResearchAssessment?