Accepted Papers & Presentations

Title & Authors	Paper	Presentation
Welcome and Introduction to the Workshop. Iraklis Klampanos, Manolis Koubarakis and Antonis Troumpoukis	—	Slides
From Self-Supervision to Trustworthy EO Foundation Models. Keynote. Ioannis Papoutsis (National Technical University of Athens, Greece)	Abstract	Slides
Comovement in Geo-referenced Time Series: A Copula-Based Approach for Clustering. Alessia Benevento, Fabrizio Durante and Roberta Pappadà	Full Text	Slides
Knowledge Graph-Enhanced Retrieval-Augmented Generation for Earth Observation Data. Roxanne El Baff, Ben Schluckebier and Tobias Hecking	Full Text	Slides
Assessing Synthetic Data Quality and Model Generalization for Planetary Imagery. Clara Salditt, Karan Molaverdikhani and Barbara Ercolano	Full Text	Slides
Weather Prediction on Mars as a Multivariate Time Series Forecasting Problem. Sagar Uprety, Amel Bennaceur, Carlos Gavidia-Calderon, James Holmes, Manish Patel and Kylash Rajendran	Full Text	Slides
Explainable Spatial Modeling of Groundwater Nitrate Concentrations in the Netherlands. Iulia Capralova and Juan Cardenas-Cartagena	Full Text	Slides
DA4DTE: An Agentic System for Enhancing the Accessibility of Digital Twins of Earth. Myrto Tsokanaridou, Jakob Heinrich Hackstein, Genc Hoxha, Sergios-Anestis Kefalidis, Konstantinos Plas, Begum Demir, Manolis Koubarakis, Marco Corsi, Cristian Leoni, Giorgio Pasquali, Chiara Pratola, Simone Tilia and Nicolas Longepe	Full Text	Slides
Semi-Supervised and Multi-label Classification of Remotely Sensed Images Keynote. Sašo Džeroski (Jožef Stefan Institute, Slovenia)	Abstract	Slides
Vessel Trajectory Prediction Using Robust AIS Preprocessing and Dual-Self-Attention GRU. Marilena Sinni and Dimitris M. Kyriazanos	Full Text	Slides
LLM-Driven Knowledge Graph Construction from Earth Observation Data for Extreme Events. Theodoros Aivalis, Iraklis A. Klampanos and Antonis Troumpoukis	Full Text	Slides
Elevation Data Integration Approaches for Deep Learning-Based 2-m Temperature Downscaling. Antigoni Moira, Stelios Karozis, Theodoros Giannakopoulos, Effrosyni Karakitsou, Nikolaos Gounaris and Athanasios Sfetsos	Full Text	Slides
An Explainable Multi-Source Unsupervised Domain Adaptation Framework Using Contrastive Learning and Adaptive Clustering for Remote Sensing Scene Classification. Binu Jose, Pranesh Das, Ebrahim Ghaderpour and Paolo Mazzanti	Full Text	Slides
A Deep Learning Approach to Evaluate Individual Predictors for Extreme Precipitation in Greece. Vasileios Vatellis, Stelios Karozis, Iraklis Klampanos, Antonis Troumpoukis and Antonis Gkanios	Full Text	Slides

☞ DARES'25 Programme

At a glance

What:	Full-day workshop
When:	25 October 2025
Where:	Bologna, Italy

Description

The Earth and Space Sciences are characterised by an exponential growth in data volume and complexity, and are derived from diverse sources such as satellite imagery, in-situ measurements, and simulations. Alongside AI, Europe invests heavily on Earth Observation and Space Programmes. Through Copernicus, a flagship initiative of the European Union, as well as via considerable and sustained contributions towards EUMETSAT, ESA, ECMWF and elsewhere, Europe has become a hub for high quality data and services. These programs, supported by substantial financial and technological resources, provide a crucial foundation for advancing scientific understanding, addressing global challenges and fueling innovation. This abundance of data presents both unprecedented opportunities for scientific discovery and significant challenges in data management, analysis, and interpretation.

This workshop aims to bring together researchers exploring how AI methods, including machine learning, symbolic approaches, and their intersection, can be effectively applied to transform, explain, fuse, or otherwise prepare large and complex datasets from diverse sources within Earth & Space Sciences for use and reuse in downstream applications. In addition, the workshop welcomes contributions to AI-powered semantic data annotation and metadata enrichment, due to their significance in improving data discoverability, understandability, and reusability.

List of relevant topic areas:

A non-exhaustive list of topics relevant to this session is the following:

• Data Preprocessing and Enhancement: AI-driven techniques for preprocessing and enhancing diverse datasets, including noise reduction, feature extraction, and data augmentation.
• Data Fusion and Integration: Address the challenges and solutions related to integrating data from multiple sources, including satellite, aerial, and ground-based sensors, using AI-based fusion techniques.
• Time-Series Analysis: AI models and approaches for analysing time-series data from various sources in Earth & Space Sciences.
• Data Reusability and Open Data Initiatives: AI for complementing best practices for making Earth & Space Science data more accessible, reusable, and interoperable.
• Semantic Data Annotation: AI-driven semantic annotation, metadata enrichment, and knowledge extraction to improve data discovery and understanding in Earth & Space Sciences.
• Data Quality Assessment: AI-driven data quality assessment, anomaly detection, and error correction in Earth & Space Science datasets.
• AI-Enhanced Data Products: AI-powered data products and services that leverage Earth & Space Science data for various sectors.
• AI for Data Search and Retrieval: Discuss AI algorithms and technologies that enhance the search and retrieval of relevant Earth & Space Science data from large archives.
• Data Privacy and Ethics: Ethical and privacy considerations when applying AI to generate Earth & Space Science data derivatives, including issues related to personal privacy, consent, and data security.

Workshop organizers

	Iraklis A. Klampanos, University of Glasgow, UK Email: What?
	Antonis Troumpoukis, National Centre for Scientific Research "Demokritos", Greece
	Manolis Koubarakis, National and Kapodistrian University of Athens, Greece

Important dates

July 30, 2025:	DARES'25 Submissions due (extended)
August 22, 2025:	DARES'25 Notification of acceptance
August 29, 2025:	Preliminary DARES'25 Programme w/ keynote speakers
September 12, 2025:	Camera-ready papers due for accepted papers
October 25, 2025:	DARES'25 Workshop takes place, as part of the ECAI'25 Conference

Submission information

We welcome:

• full papers (8-12 pages), and
• extended abstracts and short articles (2-6 pages)

The proceedings of the Workshop will be published as CEUR Workshop Proceedings (CEUR-WS.org).

All contributions will undergo single-blind review and should be formatted in the single-column CEURART style, according to CEUR-WS requirements.

Selected contributions may be published as a Special Issue in an International Journal, depending on participation and quality. Confirmation and more details will be provided at a later date.

Program Committee

Mohanad Albughdadi, European Center for Medium-Range Weather Forecasts (ECMWF), Germany
Vasileios Baousis, European Center for Medium-Range Weather Forecasts (ECMWF), UK
Omar Barrilero, European Union Satellite Centre (SATCEN), Spain
Lorenzo Bruzzone, University of Trento, Italy
Mihai Datcu, German Aerospace Center (DLR), Germany
Fani Deligianni, University of Glasgow, UK
Begüm Demir, Technical University of Berlin, Germany
Zoe Falomir, Umeå University, Sweden
Sandro Fiore, University of Trento, Italy
Stelios Karozis, NCSR “Demokritos”, Greece
Spyros Mouzakitis, National Technical University of Athens, Greece
Christian Pagé, CERFACS, France
Ioannis Papoutsis, National Technical University of Athens, Greece
Claudia Paris, University of Twente, The Netherlands
Dimitrios Skoutas, Athena Research Center, Greece
Rafael Tolosana-Calasanz, University of Zaragoza, Spain