ONE PAPER. ONE INSIGHT. #08 ➡️ A trained model can teach us more than its predictions. Explainable ML uncovered data gaps, redundancy, and unreliable predictions in an air quality dataset. Why it matters: Every prediction is an opportunity to learn something about the data. Stadtler et al., 2022

ONE PAPER. ONE INSIGHT. #07 ➡️ The most useful synthetic data fills the model’s information gaps-not necessarily the ones we expect. Subtle cues like shadows improved detection more than expected. Why it matters: Model’s blind spots can matter more than realistic synthetic data. Weber et al., 2022

ONE PAPER. ONE INSIGHT. #06 ➡️ Good models should adapt - not just predict. Our model learns how to adapt to new Earth observation tasks from only a few examples, across sensors and resolutions. Why it matters: Less data can still be enough with models that adapt quickly. Rußwurm et al., 2024

ONE PAPER. ONE INSIGHT. #05 ➡️ Machine learning and process-based models can be stronger together. We combined crop models with generative AI to simulate future crop development. Why it matters: It enables understanding and simulating complex real-world systems. Drees et al., Plant Methods, 2024

📄 ONE PAPER. ONE INSIGHT. #04 ➡️ Meaningful explanations should reflect the model, not the method. We harmonized attribution maps to obtain more consistent explanations. Why it matters: Explanations should reflect learned representations rather than explainability artifacts. Stomberg et al., 2025

ONE PAPER. ONE INSIGHT. #03 ➡️ ML is not only about optimizing models, but also about optimizing data. Why it matters: Data-centric ML turns data improvement from a tedious preprocessing step into a scientific problem that can be studied, optimized, and automated. Roscher et al. 2024

ONE PAPER. ONE INSIGHT. #02 A prediction without uncertainty is only half an answer. Instead of one answer, conformal prediction can provide a set of plausible answers when uncertainty is high. Why it matters: An honest “I don’t know” can be better than overconfident mistakes. Farag et al. 2025

ONE PAPER. ONE INSIGHT. #01 Models might learn shortcuts instead of biology. Explainability revealed that our cauliflower harvest-readiness models sometimes relied on cues not related to the underlying biology. Why it matters: A model can be right for the wrong reason. 📄 Kierdorf et al., 2023

Last October, I started my new professorship in Machine Learning in Agriculture at the University of Bonn. Our research explores the intersection of Earth observation, robotics, and machine learning. 🧵⬇️

🛰🌍 Why B03 matters: B03 PI @ribana.bsky.social leads #CRC1502 DETECT using deep learning with Sentinel-2/Landsat data to classify land cover from sparse labels, supporting uncertainty-aware EO products for climate and land-surface research. 🎥 sfb1502.de/resources/vi... 🌎 sfb1502.de/projects/clu...

📝 Working on your submission for GCPR 2026? Final stretch! Regular paper submissions are due in less than 3 days: End of May 29, 2026 (CEST)! 🚀 We’re excited to see your cutting-edge work. Good luck with the final edits!

🎉 7 new ELLIS Units join our network Units Franconia 🇩🇪, Grenoble 🇫🇷, and NRW 🇩🇪 strengthen existing hubs; Units Slovenia 🇸🇮and Sweden 🇸🇪 add new countries; Units Czechia 🇨🇿& Denmark 🇩🇰 expand from city to national Units. Learn more: https://bit.ly/4nN3vcD

📢 Call for Papers: #VMV2026 #GCPR2026 is joined by the International Symposium on Vision, Modeling, and Visualization! We invite high-quality submissions across all areas of visual computing. We are excited to see your work! 👉 Learn more & submit: www.gcpr-vmv.de/year/2026/vmv

Exciting news! The DAGM German Conference on Pattern Recognition #GCPR2026 will be hosted together with the Vision, Modeling, and Visualization #VMV2026 Conference at the University of Siegen on September 22-25, so mark your calendars! 📅 More info: www.gcpr-vmv.de/year/2026

Are you a researcher in the field of photogrammetry 📷 and remote sensing 🌍 ? We cordially invite you to submit a paper to our Photogrammetry & Remote Sensing Special Track @ German Conference on Pattern Recognition (GCPR). The deadline for regular papers is May 21! www.gcpr-vmv.de/year/2026

🛰️ Our #CRC1502 #DETECT member Prof. @ribana.bsky.social joins #Kinderuni @unibonn.bsky.social with: “Smart computers for nature—how satellites & robots help us understand and protect our environment.” Mon 26 Jan 2026, 17:15–18:00 (Bonn). Free entry (8–13 years old). Info: www.uni-bonn.de/kinderuni

🏹 Job alert: Plant Phenology, Remote Sensing, Causal Modeling at University of Jena 📍 Jena 🇩🇪 📅 Apply by Jan 3rd 🔗 https://bit.ly/44kYQ9n

🌍What would my home region look like in a different climate? ClimSat is a diffusion autoencoder for climate-conditional satellite image generation by a #CRC1502 member. It can simulate how landscapes might appear under different climatic conditions. 📜 doi.org/10.1016/j.sr... 🎥 youtu.be/YAPaGKBGJ3g

ClimSat - Climate-conditional Satellite Image Editing with Diffusion Autoencoders

YouTube video by Detect-Content-Creator

youtu.be

💻Check out the new publication from our #CRC1502 members Johannes Leonhardt and @ribana.bsky.social about Conformal Prediction. It has been found to give statistically valid uncertainty for ML in digital agriculture. It supports crop decisions with calibrated intervals. 📜 doi.org/10.1016/j.co...

📢 CALL FOR PAPERS: Closing soon! Connecting Data-Driven and Physical Approaches: Application to Climate Modeling and Earth System Observation A special collection building upon a workshop at #EGU25. ⏰ 31 October 2025 ℹ️ https://bit.ly/4k09cBu #climate #AI #forecasting

We welcome your contributions to our Thematic Session "Data-Centric Learning for Geospatial Data" at ISPRS 2026, Toronto 4-11 July, 2026: www.isprs2026toronto.com/thematic-ses... ❗️Submission deadline is November 3, 2025 🐙Organizers: Devis Tuia, Maria Vakalopoulou, Ribana Roscher

ISPRS 2026 Thematic Sessions

Explore ISPRS 2026 Thematic Sessions – structured presentations and discussions organized around the Congress’s core themes and scientific priorities.

isprs2026toronto.com

We have a PhD opening in Berlin on "Responsible Data Engineering", with a focus on data preparation for ML/AI systems. This is a fully-funded position with salary level E13 at the newly founded DEEM Lab, as part of @bifold.berlin . Details available at deem.berlin#jobs-2225

Are you a researcher in photogrammetry 📷 and remote sensing 🌍? We cordially invite you to submit a paper to our Photogrammetry & Remote Sensing Special Track @ German Conference on Pattern Recognition (GCPR). The deadline is June 04! More info: dagm-gcpr.de/year/2025

Subscribe to receive the next ones directly in your inbox! uni-bonn.us6.list-manage.com/subscribe?u=...

IAPR TC7 has changed its name to Earth Observation! Subscribe to the newsletter for news on upcoming conferences and deadlines: uni-bonn.us6.list-manage.com/subscribe?u=...

The new newsletter, with future deadlines and event is out! mailchi.mp/f4662e4d0c6c...

🚨 FEW DAYS LEFT! Abstract submission for #HAICON25 closes March 17! Join us June 3–5 in Karlsruhe for 3 days of AI-driven innovation, featuring keynotes from @mmbronstein.bsky.social & @ribana.bsky.social . 📢 Submit now: haicon.cc/contributions/ 🔗 Learn more: haicon.cc

🚨 Are you a student passionate about innovation, technology, and building connections? This is your chance to attend leading events connecting female researchers in computer science and engineering 🧳

FSTA_2025

Are you a student passionate about innovation, technology, and building connections? This is your chance to attend leading events connecting female researchers in computer science and engineering—we…

buff.ly

Get your Master's degree in "Mobile Robotics" from The University of Bonn and become one of tomorrow's innovators in robotics and intelligent systems. The program starts in the fall of this year. See: www.moro.uni-bonn.de