Strong performance on clean benchmarks is not enough. Vision systems must also survive blur, noise, compression, weather, and other real-world shifts. Introducing RoCo-Spring, an official @NeurIPSConf 2026 Competition Track challenge on robust dense correspondence. Compete in: Optical Flow | Stereo Matching | Scene Flow | Exploration We have made participation easy: register your team, download the data, clone the PTLFlow-based starter devkit, run the pipelines, and submit. 🌐 roco-spring.github.io/ 💻 github.com/hmorimitsu/roco-s… Questions? [email protected] or DM me. #NeurIPS2026 #ComputerVision #Robustness Organizers: @shashankska, @JennySchmalfuss, Victor Oei, Katrin Bauer, Henrique Morimitsu, Andres Bruhn, @margret_keuper Affiliations: @dwsunima, @nvidia, @cvml_mpiinf, @Uni_Stuttgart, and University of Science and Technology Beijing.

And @UniSiegen

Work in collaboration with @margret_keuper , @cvml_mpiinf and @dwsunima !

🚨HiWi position @dwsunima! Join the SpatialBenchRAG project to work with #LLMs, #RAG & multimodal data + contribute to open science & publications 🚀 ⏰Starts: Nov 2025 🔗 Details: uni-mannheim.de/dws/career/h… Please RT!

Our model AIM improves interpretability by learning to mask out irrelevant features in a self-supervised way. Joint work @cvml_mpiinf and @dwsunima !

New openings. Topic: social bias detection+analysis with LLMs across time (1950-now) & languages. There are 2 Post-Doc/PhD positions, supervised by@egere14 (@utn_nuremberg)+Simone Ponzetto (@dwsunima). Fully funded, up to 3 yrs. More infos: nl2g.github.io/positions

📢📢👇New job openings. Topic: social bias detection+analysis with LLMs across time (1950-now) & languages. There are 2 Post-Doc/PhD positions, supervised by @egere14 (@utn_nuremberg)+Simone Ponzetto (@dwsunima). Fully funded, up to 3 yrs. More infos: nl2g.github.io/positions

Continued....Announcing our Publicity Chairs and Local Organizer: Publicity Chairs: Shashank Agnihotri and Adeel Pervez @shashankska @dwsunima @ISTAustria Local Organizer: Julia Willi @CVisionFreiburg

Thanks to its walk-based corpus generation, Walk&Retrieve is: ✅ Adaptable to dynamic KGs ✅ Efficient: no fine-tuning of backbone LLM, single LLM call/query ✅ Compatible w/ any LLM Achieving: 📈More accurate answers 📉Fewer hallucinated or missing responses ⏳Low query latency

📢 Introducing Walk&Retrieve, a simple yet effective zero-shot #RAG framework based on #knowledgegraph walks! Arxiv : arxiv.org/abs/2505.16849 GitHub: github.com/MartinBoeckling/K… Joint work w/ Martin Böckling @heikopaulheim @dwsunima @ir_rag_sigir #SIGIR2025 Details 👇

If you are at #wacv2025 and interested in fair adversarial training, don't miss our poster presented by first author @TEJASWINIMEDI13 ! Joint work @dwsunima and @cvml_mpiinf

Amazing work! Very interesting 😄

I spy with my little eye... a WACV paper. We investigate visual frame detection using Minimum Cost Multicut clustering. Check out the paper to see which foundation model embedding space works best: arxiv.org/pdf/2412.01296 @IsaacBr45419303 @margret_keuper @SteWalter @dwsunima

Paper: openreview.net/forum?id=iVMc… (Camera-Ready coming soon) Code & Data: github.com/paulgavrikov/vlm_… Work by me, @jovita_lukasik, @jung_vision, Robert Geirhos, @jmie_mirza, @margret_keuper, @JanisKeuper. @dwsunima, @cvml_mpiinf, @MIT_CSAIL, @UniSiegen, @GoogleDeepMind

We're hiring! Want to join @dwsunima for a one-year PostDoc in an exciting project on the crossroads of #knowledgegraph research and #medical applications? Apply now: uni-mannheim.de/dws/career/h…

💪We train NaSE on 2 curated massively #multilingual news datasets, perfect for #NLProc tasks!🔥 👉 Explore on @huggingface🤗: 📁PolyNews: huggingface.co/datasets/aian… 📁PolyNewsParallel: huggingface.co/datasets/aian… Joint work w/ @fdschmidt @gg42554 @heikopaulheim @dwsunima

We are excited to hear @SwetaMahajan1 speak about her recent work on XAI, today live at Uni Mannheim. @dwsunima @KeuperLabs @margret_keuper

Huu Tan Mai presenting joint work of @dwsunima and @BoschGlobal research at #iswc2024, exploring the question whether #LLMs can adapt to unseen domains. Paper: arxiv.org/abs/2407.19998

@dwsunima's @svenhertling presenting amazing work on #KnowledgeGraph construction w/ #LLMs and function calls at @lm_kbc #iswc2024. Joint work w/ @lysander07 from @fiziseka. Paper: lm-kbc.github.io/workshop202…

Interesting new work by @PaulGavrikov in collaboration with @margret_keuper @JanisKeuper @KeuperLabs @cvml_mpiinf @dwsunima @shashankska that opens new possibilities for future work towards interpretable Deep Learning!