News & Updates

Recent activities, publications, and events

30
Aug
2026
Publication

Evolutionary Soups Accepted at EMNLP 2026

Our paper "Evolutionary Soups: Evolving Mixture-of-Experts for Multi-Objective LLM Alignment" has been accepted to the EMNLP 2026 Main Conference. The work introduces an evolutionary mixture-of-experts approach for controllable multi-objective alignment of large language models.
Learn more
14
Aug
2026
Publication

Paper Accepted at ICPR 2026

Our paper "Enhancing RL Generalizability in Robotics through SHAP Analysis of Algorithms and Hyperparameters" has been accepted at ICPR 2026. The work studies how reinforcement-learning algorithms and hyperparameters influence cross-environment generalization using SHAP-based explanations.
Learn more
07
Aug
2026
Publication

Two FAIR Digital Object Papers Accepted at ISWC and CIKM 2026

Two new papers advancing FAIR Digital Objects have been accepted at major knowledge-representation and information-management venues. "Autonomous FAIR Digital Objects: From Passive Assertions to Active Knowledge" was accepted at ISWC 2026, while "Automated Construction of FAIR Digital Object Knowledge Graphs from Flat Cultural Heritage Records" was accepted as a short paper at CIKM 2026. The latter applies FAIR Digital Objects and CIDOC-CRM to cultural heritage records from Europeana as part of the ARXIVE-ECCCH project.
Learn more
07
Jul
2026
Achievement

New DFG-Funded Project on Synthetic Health Data

Our project "Synthetic Data as a Bridge Across Health Data Silos: Evaluating Privacy, Fidelity and Utility for Cross-Domain Patient Journey Analysis" has been approved for funding by the Deutsche Forschungsgemeinschaft (DFG) as part of its Exploration Phase. The project investigates how synthetic health data can bridge fragmented healthcare data environments while preserving privacy and supporting meaningful cross-domain patient journey analysis. The work will evaluate synthetic data along three key dimensions: privacy, fidelity, and utility, in collaboration with the Institute for Biomedical Informatics in Cologne and Berlin University of Applied Sciences.
Learn more
17
Jun
2026
Publication

TecoPrompt Accepted at ECCV 2026

Our paper "TecoPrompt: Temporal-Conservative Prompt Learning for Vision-Language Models" has been accepted at ECCV 2026. The work investigates temporal-conservative prompt learning for improving adaptation in vision-language models.
Learn more
16
Jun
2026
Publication

Two Papers at ICMR 2026

Two of our computer-vision papers appear at ICMR 2026 in Amsterdam: "MoiréNet: Leveraging Directional Priors for Compact Dual-Domain Image Demoiréing" and "TrinitySeg: A Benchmark and Baseline for Monochrome NIR Smoke Segmentation." The works address image demoiréing and near-infrared smoke segmentation, respectively.
Learn more
08
Jun
2026
Award

InstaDA Receives CVPR 2026 Compute Transparency Champion Recognition

Our CVPR 2026 paper "InstaDA: Augmenting Instance Segmentation Data with Dual-Agent System" received the Compute Transparency Champion recognition under CVPR 2026's Compute Reporting Initiative. The work introduces a dual-agent framework combining large language models and diffusion models for instance-segmentation data augmentation.
Learn more
26
May
2026
Event

Presenting Cultural Heritage Research at Digital Heritage Summit 2026

Participated in the Digital Heritage Summit 2026 in Limassol, Cyprus, where our work on FAIR Digital Objects, trustworthy AI, multimodal data processing, and cultural heritage preservation was presented. The summit brought together researchers, cultural heritage institutions, policymakers, and European initiatives working on digital heritage infrastructures.
Learn more
30
Apr
2026
Event

First In-Person AI-BioDynamics Consortium Meeting

We welcomed our German and Korean AI-BioDynamics partners to Fraunhofer FIT at Schloss Birlinghoven for the project's first in-person consortium meeting. The working session focused on current progress and upcoming work combining AI, remote sensing, and biodiversity data for scalable biodiversity monitoring.
Learn more
01
Dec
2025
Event

AI-BioDynamics Project Kicks Off!

Our German-Korean AI-BioDynamics project has officially launched. Supported by Germany's BMFTR and Korea's MSIT with an overall budget of approximately €1.3 million, the project will develop an intelligent platform combining satellite data, explainable AI, and process mining for biodiversity monitoring, conservation planning, and related decision support. The project runs through November 2028.
Learn more
10
Nov
2025
Publication

Fair-Way Presented at CIKM 2025

Our paper "FAIR Data Assessment Using LLMs: The Fair-Way" appears at the 34th ACM International Conference on Information and Knowledge Management (CIKM 2025) in Seoul. The work investigates the use of large language models to support assessment of FAIR research data.
Learn more
08
Nov
2025
Publication

Paper Accepted at AAAI 2026

Our paper "Unleashing Semantic and Geometric Priors for 3D Scene Completion" has been accepted at the 40th AAAI Conference on Artificial Intelligence. The work explores how semantic and geometric priors can improve 3D scene completion models.
Learn more
25
Aug
2025
Publication

Two Papers at SIGDIAL 2025

Two of our papers were accepted at SIGDIAL 2025 in Avignon, France. "EMORL: Ensemble Multi-Objective Reinforcement Learning for Efficient and Flexible LLM Fine-Tuning" introduces an efficient multi-objective RL framework, while "Exploring the Limits of Model Compression in LLMs" investigates knowledge distillation for question-answering tasks.
Learn more
30
Jun
2025
Publication

Q-Chain at ICMR 2025

Our paper "Q-Chain: A Causal-Aware Framework for Structural and Educational Question Generation" was presented at ICMR 2025 in Chicago. Q-Chain introduces causal reasoning into automatic question generation to improve question structure and educational value.
Learn more
23
Jun
2025
Publication

ELMTEX at AIME 2025

Our paper "ELMTEX: Fine-tuning Large Language Models for Structured Clinical Information Extraction — A Case Study on Clinical Reports" appears at AIME 2025 in Pavia, Italy. The work explores fine-tuning LLMs for structured information extraction from clinical reports.
Learn more
02
Dec
2023
Publication

Two NFDI4DS Papers Published at INFORMATIK 2023

Two collaborative NFDI4DataScience papers were published in the INFORMATIK 2023 proceedings: "NFDI4DS Infrastructure and Services" and "Research Knowledge Graphs in NFDI4DS." The papers present key components of the NFDI4DS infrastructure and its approach to organizing FAIR research artifacts through research knowledge graphs.
Learn more
01
Nov
2023
Publication

New Paper in the International Journal of Computer Vision

Our paper "Skeleton Ground Truth Extraction: Methodology, Annotation Tool and Benchmarks" was published online in the International Journal of Computer Vision. The paper introduces a methodology and annotation tool for creating consistent skeleton ground truth across shape and image datasets.
Learn more
13
Sep
2023
Talk

Presenting FAIR Digital Object Research at CoRDI 2023

Presented our work on FAIR Digital Objects and research reproducibility at the first Conference on Research Data Infrastructure (CoRDI 2023). The contribution "Enhancing Reproducibility in Research through FAIR Digital Objects" discussed how research artifacts can be made more reusable and machine-actionable.
Learn more
28
Oct
2022
Talk

Panel Discussion at FDO2022

Participated in a panel discussion at FDO2022 on employing FAIR Digital Objects to make data-driven ecosystems FAIR, with a focus on their use within NFDI4DataScience and related research-data infrastructures.
Learn more
21
Sep
2022
Award

Best Paper Award at TPDL 2022

Our paper "Whois? Deep Author Name Disambiguation using Bibliographic Data" received the Best Paper Award at the International Conference on Theory and Practice of Digital Libraries (TPDL 2022) in Padua, Italy.
Learn more

Archive

2026 2025 2023 2022