Building an Inference Engine Using AI and The World’s Largest Meta- Analysis: Lesson Learned
Bibliographic record
Abstract
This symposium describes the lessons learned during the creation of the world's largest meta-analysis, that is over 2,500 studies assessing over 160 constructs from the field of Organizational Justice. By combining cloud based meta-analytic databases, online statistical engines, curated selection of academic articles and the new LLMs, we have elevated this into an effective inference engine. Using the LLM, questions are translated into variables and specific search terms. Based on these variables, the relevant empirical results are drawn from the cloud based meta-analytic database and analyzed by an online statistical engine. Interpretation of results are enhanced by a core base of review articles as well as relevant articles drawn using the search terms from the meta-analysis database, whereupon the LLM provides custom, empirically sourced answers (with appropriate citations) back to the user in seconds. For more sophisticated queries, the LLM can create the specific R code to analyze the database, which is then executed by the statistical engine. As we will demonstrate, inference engines based on meta-analytic databases appear to be the ideal vehicle for interacting with scientific knowledge.
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 | 0.135 | 0.270 |
| Meta-epidemiology (narrow) | 0.002 | 0.003 |
| Meta-epidemiology (broad) | 0.004 | 0.009 |
| Bibliometrics | 0.011 | 0.008 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.017 | 0.018 |
| Open science | 0.005 | 0.007 |
| Research integrity | 0.003 | 0.010 |
| Insufficient payload (model declined to judge) | 0.007 | 0.004 |
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".