Pathway Logic Extended with Information Assembled from Data Extracted Selectively (PLEIADES)
Abstract
This report details our work under DARPAs Big Mechanism program, which pursued automated methods for the creation and extension of sophisticated mechanistic models through the ingestion of textual information. Under Big Mechanism, this problem was factored into three topic areas: reading, or the automated conversion of human-language text to structured semantic form; assembly, or the incorporation of reading outputs into working mechanistic models; and explanation, or the use of these enhanced models to solve problems or provide insights to humans. The program nominated a challenging problem domain, the cellular signaling pathways associated with a critical protein family called Ras, the malfunction of which is implicated in a number of types of cancer. This problem domain has a number of characteristics that made it useful to the program: the relevant mechanisms are poorly understood, the pathways in question are large, textual material treating these pathways are voluminous, and any successes in meeting the programs challenge would potentially yield immediate medical benefit.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 03, 2019
- Accession Number
- AD1081335
Entities
People
- Andrew Silberfarb
- Carolyn Talcott
- Dayne Freitag
- Mark Craven
- Richard Rohwer
Organizations
- SRI International