Applying LLMs and Semantic Technologies for Data Extraction in Literature Reviews
Bibliographic record
Abstract
This pilot study evaluates the capabilities of two LLMs, Mistral Small 3.1 and GPT-4o mini, in performing ontology-based data extraction to support literature reviews in library and information science (LIS). A sample of four published systematic reviews was selected as ground truth data. The open-access publications included in these reviews (n = 47) were collected as inputs for the models to perform semantic information extraction, using classes from the Document Components Ontology (DoCO). These preliminary findings highlight the opportunities and challenges of using AI and semantic technologies to streamline literature reviews in the social sciences. Application des GML et des technologies sémantiques pour l'extraction de données dans les revues de littérature : Une étude pilote en sciences de l'information RésuméCette étude pilote évalue les capacités de deux GML, Mistral Small 3.1 et GPT-4o mini, pour effectuer une extraction de données basée sur une ontologie pour supporter les revues de littérature en bibliothéconomie et sciences de l'information (BSI). Un échantillon de quatre revues systématiques publiées a été sélectionné comme données véridiques de base. Les publications à accès libre incluses dans ces revues (n = 47) ont été choisies comme entrées dans les modèles, pour qu'ils effectuent une extraction d'information sémantique en utilisant les catégories du Document sur les composantes de l'ontologie (DoCO). Ces résultats préliminaires soulignent les opportunités et les défis de l'utilisation de l'IA et des technologies sémantiques pour l'organisation des revues littéraires en sciences sociales. Mots-ClésGML; Technologie sémantique; Extraction d'information; Synthèse de connaissances
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.059 | 0.198 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.004 |
| Bibliometrics | 0.018 | 0.014 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.007 | 0.008 |
| Open science | 0.002 | 0.005 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".