Ontologer/ORACLE-ClinicalTrials-SuccessProb-v1
Text Classification ā¢ Updated
ā¢ 8
Ontologer
company
Verified
AI & ML interests
None defined yet.
Recent Activity
Organization Card
Ontologer
Medical intelligence infrastructure. Ontologer builds structured databases, purpose-built models, and workflows for clinical trials, regulatory submissions, publications, and adverse events.
Our models are trained on data drawn from a cross-linked medical data platform spanning 80M+ records across 20+ primary-source databases.
ā ontologer.com
Data Infrastructure
80M+ structured records across 20+ primary-source databases, normalized and cross-linked.
| Category | Records |
|---|---|
| Clinical trials | 560K+ |
| Biomedical publications | 28M+ |
| Drug bioactivity & compound data | 27M+ |
| Industry payments to healthcare professionals | 16M+ |
| Healthcare provider registry | 9M+ |
| Drug & device safety reports | 350K+ |
| Regulatory clearances & filings | 174K+ |
| Medical device registry | 100K+ |
| Medical classification codes (diagnoses, procedures, drugs) | 285K+ |
| Regulatory text | 8K+ |
| Patents & IP filings | 10K+ |
Models
| Model | What it does |
|---|---|
| MOAt-ClinicalTrials-MoA-256 | Mechanism of action similarity embeddings |
| PACT-ClinicalTrials-Pop-256 | Patient population & condition similarity embeddings |
| FATE-ClinicalTrials-Outcome-256 | Historical outcome pattern similarity embeddings |
| ORACLE-ClinicalTrials-SuccessProb-v1 | Calibrated trial success probability (ROC AUC 0.734) |
Pipeline
Raw Trial ā [Ontologer Extraction Model] ā MOAt (mechanism space, 256d)
ā PACT (population space, 256d)
ā FATE (outcome pattern space, 256d)
ā ORACLE (P(success) from all three)
Access
Model weights are not publicly available. For API access, research collaboration, or enterprise inquiries, visit ontologer.com.
datasets 0
None public yet