CoopIS-ERA 2026 Abstracts


Area 1 - Early Research Achievement

Full Papers
Paper Nr: 5
Title:

Integration of Temporal Reasoning in Spacecraft Design

Authors:

Julian Lechner, Philipp Chrszon and Andreas Gerndt

Abstract: In the early phases of spacecraft design, feasibility studies are conducted to assess whether a given mission is achievable within a set of constraints. To reduce design time, feasibility studies are typically performed concurrently, meaning multiple systems engineers model the system simultaneously. Over the past few decades, various cooperative information systems such as Virtual Satellite have been developed to support systems engineers in this process. By now, the system of interest is modeled structurally, meaning it represents the system at a specific time point. In reality, however, designing and building the system is a cooperative process involving multiple stakeholders from different domains whose activities are subject to temporal constraints. In later phases, modeling the system’s temporal behavior becomes important as well. Currently, such temporal information is modeled implicitly, making it hard to check in advance whether all temporal constraints can be satisfied. This paper investigates how temporal reasoning can be integrated into a cooperative information system and contributes an extension of Virtual Satellite that enables the explicit modeling and verification of temporal constraints. We report on the lessons learned from that process to help readers with similar integrations. While our integration demonstrates the technical feasibility, the lack of accurate temporal data highlights the need to adapt modeling processes accordingly.

Paper Nr: 7
Title:

Exploiting Weak Signals in Family Conversations as a Lever to Augment Structured Data in a Caregiver Setting

Authors:

Paul Cariou, Kaoutar Akhsass, Luiz Angelo Steffenel and Manuele Kirsch Pinheiro

Abstract: Medical records typically focus on biological markers and acute events, often overlooking subtle signals from daily life (sleep, mood, feelings, treatment side effects). These “invisible” data, essential for adjusting care, are often shared informally among relatives, representing a significant source of data in home care context. In this paper, we intend to explore such data, using Large Language Models (LLMs) to structure and make them actionable for care teams. We propose to explore a pre-trained LLM in order to identify weak signals, notably potential symptoms mentioned in such informal messages. We present our first results, obtained with a corpus of about 2,000 verbatims and evaluated thanks to a LLM-as-a-judge approach.

Paper Nr: 11
Title:

Towards a Machine Learning-Supported Cooperative Information System for Multi-Index Decision Support in Sustainable Agriculture

Authors:

Piotr Robert Szczerbiak and Olga Siedlecka-Lamch

Abstract: Precision agriculture increasingly depends on satellite images, vegetation indicators, and machine learning techniques. Nevertheless, numerous current systems display analytical outcomes as distinct maps, imposing on farmers the responsibility of understanding various indicators and converting them into field-level decisions. This study introduces Filskane, an early-stage machine learning-supported cooperative information system designed for multi-index decision support in sustainable agriculture. The system integrates spatial raster data management, contextual data gathering, vegetation index analysis, and density-based clustering to convert remote sensing results into comprehensible intervention zones. The innovation of the study is not in the introduction of a new vegetation index or clustering method, but in the implementation of multi-index analysis as a collaborative decision-making framework that includes farmers, analytical services, and agricultural consulting specialists. We present the prototype utilizing satellite data and outline an evaluation strategy emphasizing interpretability, agronomic plausibility, farmer use, and technological feasibility. This paper presents an initial framework and research plan for human-centered, machine learning-supported information systems in sustainable agriculture. At this stage, the results illustrate the feasibility of the approach and require further expert and field validation.

Paper Nr: 14
Title:

Learning to Guide Ontology-Driven Geological Interpretation with Graph Neural Networks

Authors:

Aya Attia, Gautier Laurent and Imadeddine Laouici

Abstract: Building 3D geological models from sparse observations relies on expert interpretation whose reasoning is rarely documented, making models difficult to validate and reproduce. An ontology-based system, built around the Minimal Ontology for Geological Interpretation (MOGI) and a Geological Interpretation Process (GIP), was recently proposed to make this reasoning explicit. It formalizes interpretation as an iterative cycle and records every reasoning step as a trace. However, two decisions of this cycle, which situations to interpret next and which geological explanation to assign to them, are currently made randomly or with simple heuristics. We propose to learn these decisions from past reasoning traces, each trace is encoded as a graph labelled with a quality score. A Relational Graph Convolutional Network (R-GCN)-based evaluator is trained to predict this score and guide decisions during the interpretation cycle. In a proof-of-concept on a synthetic fold dataset, the evaluator ranks interpretation states with a Pearson correlation of 0.98.

Short Papers
Paper Nr: 6
Title:

Toward AI-Augmented Cooperative Engineering Workflows: Requirements and Architecture from the European Rover Challenge

Authors:

Ahmed R. Sadik, Frank Joublin, Mariusz Bujny, Antonello Ceravola and Joan Smith

Abstract: Artificial Intelligence (AI) is increasingly used in engineering, but its adoption remains concentrated on isolated activities such as coding, documentation, and information retrieval. Less attention has been given to cooperative workflows, where teams must coordinate requirements, tasks, communication, knowledge, and subsystem integration. We investigate this problem through a role-adaptive 40-question survey of European Rover Challenge (ERC) teams, with 104 responses from 14 teams. The results reveal recurring bottlenecks: difficult knowledge transfer, informal task assignment and monitoring, fragmented communication, manual requirement tracking, and substantial rework caused by late discovery of integration problems. Although 78.8% of respondents already use AI tools, their strongest unmet needs concern project coordination, requirement handling, communication support, and system-level integration. We translate these findings into requirements for AI-augmented cooperative engineering workflows and propose an assistant-system architecture connecting multimodal user interfaces, service selection, specialized AI services, and external engineering tools. The contribution is an empirically grounded set of workflow requirements and an architectural direction for hybrid human--AI engineering teams.

Paper Nr: 12
Title:

Internationalization of Digital Public Infrastructures: A Reference Architecture

Authors:

Dev Shinde, Balambiga Ayappane and Srinath Srinivasa

Abstract: Digital Public Infrastructure (DPI) refers to a set of shared digital systems, services, and platforms designed for public use, forming the substrate on which state and private entities build public-facing services. DPIs are typically designed within an overarching national legal context, but when DPI services must cross national boundaries, they encounter architectural mismatches and conflicting regulatory requirements for which no systematic architectural framework currently exists. We survey existing mechanisms for cross-border sharing of sensitive utilitarian data, reposition the problem in the context of public data flows, and propose a Trust Resolution Architecture (TRA): a jurisdiction-aware extension of an existing consent-management formalism that lets sovereign DPIs establish legitimate, purpose-scoped, post-exchange-enforceable data flows without presupposing a shared legal framework or technical stack.