A Generalized Process Model of Human Action Selection and Error and its Application to Error Prediction

Frank TamborelloNational Academies of Science
Gregory TraftonUS Naval Research Laboratory

Abstract

Our model of action selection and postcompletion error in two form filling tasks extends to skip errors in a story telling task. We also discuss how it explains perseverations in one of the aforementioned form filling tasks. Finally we discuss a predictive classifier application we built from the model’s data. The classifier could allow an autonomous agent to know when it is a bad time to interrupt a human, when a human is about to err, and how to help.

Files

A Generalized Process Model of Human Action Selection and Error and its Application to Error Prediction (1 KB)



Back to Table of Contents