How Reliable Is Semantic Search in Industrial Computing Domain? A Statistical Evaluation Pipeline
Bibliographic record
Abstract
Statistical evaluation is crucial for improving semantic search, especially in highly technical domains such as industrial computing. Although qualitative human assessments are often used, they are time-consuming, expensive, and cover only a limited range of scenarios. A key challenge in conducting statistical evaluation is generating a realistic ground truth test set that accurately represents complex domain-specific terminology, including users’ persona and background knowledge. However, once this test set is established, it enables systematic experimentation to refine AI search configurations. Our study finds that increasing text volume improves search accuracy, leading to a precision increase from 77% to 82%. However, an excess of specialized terms, indicated by a relatively higher token-to-word conversion rate, can potentially weaken semantic understanding. To mitigate this, a hybrid approach that integrates semantic search with the traditional lexical method is utilized, with OpenAI embeddings to further enhance search performance.
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 | 0.057 | 0.245 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.014 | 0.009 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.007 | 0.014 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.003 |
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".