IDEAL-X information extraction system employs adaptive learning, and could be parameterized by controlled vocabularies over the given reports. The system adheres to and extends the workflow of a previous system, ASLForm.
The system is domain-agnostic and therefore can be adapted by different working environments easily.
There are no configurations or pre-existing training sets required, thus the system avoids costly engagement of annotator and linguistic experts.
The system provides an intuitive interface and interaction, and it learns covertly from users’ interactions without disruptive user engagements. This significantly simplifies the user effort.
Improvement to the decision model of the system occurs continuously and incrementally, which renders increasingly accurate output optimized for the targeted clinical application.