D3.7 Technical report on information extraction from heterogeneous data using TDM
Bibliographic record
Abstract
This deliverable summarises the activities within Task 3.3. that were led by CLARIN ERIC with the main partners CLARIN/Athena, CLARIN/CUNI, and DARIAH/UGOE, and additional collaboration with SciencesPo. Over the course of M1-M38 the partners developed three demonstration scenarios that highlight the value of NLP (Natural Language Processing) technologies for the SSH field and investigated which aspects of the outcomes of T3.3. and in which form can be shared via SSH Open Marketplace. Three types of scenarios include: (1) Application of TDM (Text Data Mining) to large bodies of multilingual texts on the use case of processing Collective Bargaining Agreements (CBAs); (2) Integration of linguistic analysis for information extraction into SSH tasks on the use case of verbal aggression detection in the context of social media. (3) TDM handling of heterogeneous data on the use case of processing the intertextuality phenomena in European drama history. All use cases work with data in multiple languages, and created pipelines take this multilinguality into account. The outcome of the demonstrations are stored to be accessed after the end of the project as online code notebooks and a workflow on SSH Open Marketplace (use case 1), service in the EOSC portal (use case 2), and a set of python scripts on the publicly accessible GitHub pages and a workflow on SSH Open Marketplace (use case 3).
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 | 0.016 | 0.045 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.005 | 0.005 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.009 | 0.008 |
| Open science | 0.004 | 0.010 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.056 | 0.085 |
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".