aiqbench analysis
ArticleMiner introduces a novel framework that combines language models with domain-specific human-authored modules to construct knowledge graphs from scientific publications, effectively bridging the gap between raw table data and semantic meaning. By reconciling evidence from multiple parsers and applying field-specific constraints, the system outperforms standard few-shot LLM baselines in extracting quantitative facts across diverse disciplines like materials science and geochemistry. This approach advances the state of AI-driven literature mining by ensuring that extracted numerical values are correctly contextualized according to their original scientific conventions.
ArticleMiner: Ontology-Guided Knowledge Graph Construction from Scientific Publications
arXiv:2609.25607v1 Announce Type: new Abstract: Scientific papers keep much of their quantitative content in tables and supplementary files, where a number means something only through its header, caption, unit, analytical method, and the conventions of its field. Recovering the rows and columns of a table is therefore not the same as recovering the scientific fact it reports. Most semantic table-interpretation methods assume that a clean table is already available and subsequently map its cells or columns to ontology terms, whereas most publication-level extraction systems are designed for a single domain. We study a middle path: a shared process that reads a paper and its supplementary files, gathers evidence from several parsers and a language model, and reconciles that evidence, while a bounded human-authored task module for each task supplies the domain meaning. The module lists the canonical names the graph may use, the surface forms that map to them, a small set of derivation rules and validity constraints, an identity key, and the bindings used to write RDF. It defines what a task is allowed to emit; it does not try to list every convention of a field. We build four such modules (for drug-discovery chemistry, materials science, machine learning, and mineral geochemistry) in the ArticleMiner framework, and evaluate them on 163 papers, including a new geochemistry benchmark with expert-curated ground truth. In comparisons against a same-LLM few-shot baseline, the point estimates favor ArticleMiner on all four tasks, with uncertainty on the two smaller benchmarks. The geochemistry comparison also includes access to supplementary files, so its improvement cannot be attributed to domain guidance alone.
Abstract:Scientific papers keep much of their quantitative content in tables and supplementary files, where a number means something only through its header, caption, unit, analytical method, and the conventions of its field. Recovering the rows and columns of a table is therefore not the same as recovering the scientific fact it reports. Most semantic table-interpretation methods assume that a clean table is already available and subsequently map its cells or columns to ontology terms, whereas most publication-level extraction systems are designed for a single domain. We study a middle path: a shared process that reads a paper and its supplementary files, gathers evidence from several parsers and a language model, and reconciles that evidence, while a bounded human-authored task module for each task supplies the domain meaning. The module lists the canonical names the graph may use, the surface forms that map to them, a small set of derivation rules and validity constraints, an identity key, and the bindings used to write RDF. It defines what a task is allowed to emit; it does not try to list every convention of a field. We build four such modules (for drug-discovery chemistry, materials science, machine learning, and mineral geochemistry) in the ArticleMiner framework, and evaluate them on 163 papers, including a new geochemistry benchmark with expert-curated ground truth. In comparisons against a same-LLM few-shot baseline, the point estimates favor ArticleMiner on all four tasks, with uncertainty on the two smaller benchmarks. The geochemistry comparison also includes access to supplementary files, so its improvement cannot be attributed to domain guidance alone.
Artificial Intelligence (cs.AI)
arXiv:2609.25607 [cs.AI]
(or arXiv:2609.25607v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.25607 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 353 words
Related coverage
LLM-Guided Ontology-Driven Knowledge Graph Construction from Unstructured Text
arXiv · Sep 30, 2026
Graph2Idea:Retrieval-Augmented Scientific Idea Generation with Graph-Structured Contexts
arXiv · Jun 10, 2026
Graph2Idea:Retrieval-Augmented Scientific Idea Generation with Graph-Structured Contexts
arXiv · Jun 9, 2026
Graph2Idea:Retrieval-Augmented Scientific Idea Generation with Graph-Structured Contexts
arXiv · Jun 11, 2026