MaRDMO: Future Gateway to FAIR Mathematical Data

TL;DR

MaRDMO addresses the lack of standardized management for mathematical research data by embedding it into the RDMO framework. It provides guided documentation of mathematical models, algorithms, and interdisciplinary workflows, backed by connections to MathModDB, MathAlgoDB, MaRDI Portal, and Wikidata, and enforces data quality through automated validation and export mechanisms. The approach leverages a unified data model and federated knowledge graphs to enable findable, accessible, interoperable, and reusable data, with PID-based provenance and cross-platform search. The work envisions future expansion to more MaRDI services, AI-assisted curation, and broader, cross-disciplinary adoption.

Abstract

Mathematical research data plays a crucial role across scientific disciplines, yet its documentation and dissemination remain challenging due to the lack of standardized research data management practices. The MaRDMO Plugin addresses these challenges by integrating mathematical models, algorithms, and interdisciplinary workflows into the established framework of the Research Data Management Organiser (RDMO). Built on FAIR principles, MaRDMO enables structured documentation and retrieval of mathematical research data through guided questionnaires. It connects to multiple knowledge graphs, including MathModDB, MathAlgoDB, and the MaRDI Portal. Users can document and search for models, algorithms, and workflows via dynamic selection interfaces that also leverage other sources such as Wikidata. The plugin facilitates the export to the individual MaRDI services, ensuring data quality through automated validation. By embedding mathematical research data management into the widely adopted RDMO platform, MaRDMO represents a significant step toward making mathematical research data more findable, accessible, and reusable.

Figures (3)

• Figure 1: Data model employed by the MaRDMO Plugin. Nodes represent classes, while labeled, directed edges indicate relations between them. Classes relevant to algorithms, mathematical models, and interdisciplinary workflows are depicted in red, blue, and green, respectively. Two-colored nodes signify relevance to multiple data types. The Method class is shared between interdisciplinary workflows and algorithms, but only methods classified as algorithms are considered part of the algorithmic domain. Software and Instruments are distinct classes, combined here to parallel the Tool class in Metadata4Ing; however, only Software belongs to the algorithmic domain and is linked to the Hardware class. An asterisk in the edge label denotes the existence of multiple distinct relations. The publication class relevant for algorithms, mathematical models and interdisciplinary workflows is omitted for clarity.
• Figure 2: RDMO user interface with the Methods page of the interdisciplinary workflow catalog. Uzawa Iteration is searched, corresponding items are found in MathAlgoDB, the MaRDI Portal and Wikidata.
• Figure 3: Automating schemes within the catalogs for mathematical models (a), algorithms (b) and interdisciplinary workflows (c) displaying the individual classes of the corresponding ontologies. The selection of items from either class leads to an automatic insertion of all downstream items. For clarity, the publication class, which is a downstream class of all other classes, has been omitted.