A collaborative filtering-based approach to biomedical knowledge discovery
Bibliographic record
Abstract
Motivation: The increase in publication rates makes it challenging for an individual researcher to stay abreast of all relevant research in order to find novel research hypotheses. Literature-based discovery methods make use of knowledge graphs built using text mining and can infer future associations between biomedical concepts that will likely occur in new publications. These predictions are a valuable resource for researchers to explore a research topic. Current methods for prediction are based on the local structure of the knowledge graph. A method that uses global knowledge from across the knowledge graph needs to be developed in order to make knowledge discovery a frequently used tool by researchers. Results: We propose an approach based on the singular value decomposition (SVD) that is able to combine data from across the knowledge graph through a reduced representation. Using cooccurrence data extracted from published literature, we show that SVD performs better than the leading methods for scoring discoveries. We also show the diminishing predictive power of knowledge discovery as we compare our predictions with real associations that appear further into the future. Finally, we examine the strengths and weaknesses of the SVD approach against another well-performing system using several predicted associations. Availability and implementation: All code and results files for this analysis can be accessed at https://github.com/jakelever/knowledgediscovery. Contact: sjones@bcgsc.ca. Supplementary information: Supplementary data are available at Bioinformatics online.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".