System Overview

At the highest level, OpenSherlock is a society of agents architecture. Some of its agents are:

• ASR for reading text documents for the purpose of growing and maintaining models in domains of interest
• TopicMap for hosting those models
• TSC for building Qualitative Proceee Theoretic models to add to the TopicMap, and to perform evolutionary discovery of new concepts within those models
• Literature-based Discovery (LBD) to find unconnected ideas in those models
• Deep Question Answering to serve as a kind of research interface into those models.

Architecture

Components

• Anticipatory Story Reader
• Long-term memory
• Short-term working memory