Modelpublic

Academica/AcademicaMed-PACT-ClinicalTrials-Pop-256

sourceHugging Facecc-by-nc-nd-4.0updated 2mo agoView on Hugging Face

0likes12downloads

Model Card

AcademicaMed-PACT-ClinicalTrials-Pop-256

PACT is the population and condition embedding model in the AcademicaMed clinical-trial model family.

The model represents trial text as a 256-dimensional vector. Its contrastive objective groups trials through standardized MeSH condition terms.

Model details

Property	Value
Base model	`NeuML/pubmedbert-base-embeddings`
Output dimension	256
Training method	Triplet-loss contrastive learning
Training corpus	563,845 ClinicalTrials.gov entries
Grouping signal	4,960 MeSH condition terms
Epochs	10
Batch size	32
Learning rate	`2e-5`
Loss margin	0.5
Training hardware	NVIDIA DGX Spark, GB10 Blackwell

Intended use

PACT supports similarity analysis by enrolled population and condition. It does not estimate treatment efficacy or patient-level outcomes.

Architecture

text

Trial text -> PubMedBERT embedding -> dense 384-to-256 projection -> L2-normalized vector

Limitations

The training signal depends on ClinicalTrials.gov records and MeSH condition assignments. Missing or inconsistent records can affect similarity results.

This repository contains the original published model card and configuration scope. The source repository did not publish executable weight files.

AcademicaMed family

—MOAt: mechanism-of-action similarity.
—FATE: historical outcome-pattern similarity.
—ORACLE: calibrated synthesis classifier.