search menu icon-carat-right cmu-wordmark

Human-Computer Decision Systems

October 2015 Presentation
Brian Lindauer

Describes work to use learning theory advances to account for persistent human expert teams and experiments to improve the human-computer decision system

Publisher:

Software Engineering Institute

Abstract

After examining why current approaches are inadequate, the SEI researchers examined what is needed to know whether a new approach works. They explore these factors: realistic data (class and feature distributions that relate to a transition domain), human participants (actual errors and learning patterns), and ground truth (because we know labelers are fallible).