Draft:Open Research Knowledge Graph

Open Research Knowledge Graph (ORKG) is a digital infrastructure project that aims to improve scholarly communication through structured and semantically rich representations of research contributions. It provides a collaborative platform where researchers can create, share, and interlink scholarly knowledge in a machine-readable form, thus enabling advanced services such as comparison, intelligent search, and knowledge discovery.

Overview

The ORKG is developed and maintained by the TIB – Leibniz Information Centre for Science and Technology in Germany. It represents a shift from traditional document-centric publishing towards knowledge-centric communication. By modeling research contributions as structured data, the ORKG facilitates enhanced retrieval, integration, and automated reasoning over scholarly outputs.

The project addresses the limitations of conventional scholarly articles by decomposing them into fine-grained semantic entities (e.g., problems, methods, results), which are then stored in a graph-based structure. This allows for novel ways of consuming and analyzing scientific content, such as dynamic comparisons across multiple research papers or real-time synthesis of state-of-the-art knowledge.

Features

Key features of the ORKG include:

Semantic Templates: Guided forms to help researchers describe their work using standardized vocabularies.
Contribution Comparison: Automated comparison of research contributions across multiple works.
Visualization Tools: Interactive visualizations to navigate and explore scholarly knowledge.
Persistent Identifiers: Integration with DOI and ORCID systems for reliable citation and attribution.
API Access: A public API for querying and contributing data programmatically.

Applications

The ORKG is used in various domains, including computer science, physics, engineering, and medicine. Its ability to semantically link research outputs supports tasks such as systematic reviews, meta-analyses, and the detection of emerging trends.

Researchers, data curators, and institutions use the ORKG to:

Publish machine-readable research knowledge.
Compare related contributions in a given domain.
Facilitate reproducibility by linking datasets, software, and methodologies.

ORKG Ask

ORKG Ask is a natural language interface to a vast corpus of scientific literature (covering more than 80 Million articles) and the Open Research Knowledge Graph that enables users to query scholarly knowledge using plain English questions. Instead of requiring formal query languages like SPARQL, ORKG Ask translates user questions into structured graph queries, returning answers drawn from the scholarly corpus and semantically annotated content in the ORKG.

This feature is designed to make scholarly knowledge more accessible to a broader audience, including researchers unfamiliar with query syntax, students, policy makers, and the general public. By leveraging natural language processing (NLP) and semantic matching techniques, ORKG Ask supports intelligent retrieval of research contributions, methods, datasets, and comparisons.

Key capabilities of ORKG Ask include:

Question Answering: Users can pose questions such as "What are the methods used for COVID-19 detection in medical imaging?" and receive synthesized results across multiple contributions.
Comparison Generation: ORKG Ask automatically generates comparison tables for questions involving multiple entities or approaches.
Explanatory Outputs: The system highlights the provenance of results by linking directly to the underlying research contributions and metadata.

ORKG Ask represents a step toward conversational search in scholarly communication and aligns with the broader vision of making research results more FAIR (Findable, Accessible, Interoperable, and Reusable).

Technical Architecture

The ORKG^[1] is based on a knowledge graph architecture using RDF (Resource Description Framework) and related Semantic Web technologies. It provides a RESTful API and a SPARQL endpoint for querying the underlying data.

Its backend services are implemented using modern software stacks including Kotlin and Neo4j, with frontend components built using React.js. The platform is open source and available on GitLab under a permissive license.

Development and Community

The ORKG is an ongoing initiative of TIB and benefits from collaborations with international partners, universities, and research organizations. The project is supported by several grants from the German Research Foundation (DFG) and the European Union.

Community engagement is facilitated through workshops, hackathons, and open-source contributions. The ORKG also contributes to discussions on Open Science and FAIR data principles.

References

^ Auer, Sören; Oelen, Allard; Haris, Muhammad; Stocker, Markus; D’Souza, Jennifer; Farfar, Kheir Eddine; Vogt, Lars; Prinz, Manuel; Wiens, Vitalis; Jaradeh, Mohamad Yaser (2020). "Improving Access to Scientific Literature with Knowledge Graphs". Bibliothek Forschung und Praxis. 44 (3): 516–529. doi:10.1515/bfp-2020-2042.

Cite error: A list-defined reference named "orkgportal" is not used in the content (see the help page).
Cite error: A list-defined reference named "orkgcode" is not used in the content (see the help page).

External links

[Auer2020-1] Auer, Sören; Oelen, Allard; Haris, Muhammad; Stocker, Markus; D’Souza, Jennifer; Farfar, Kheir Eddine; Vogt, Lars; Prinz, Manuel; Wiens, Vitalis; Jaradeh, Mohamad Yaser (2020). "Improving Access to Scientific Literature with Knowledge Graphs". Bibliothek Forschung und Praxis. 44 (3): 516–529. doi:10.1515/bfp-2020-2042.

[1]