Understanding Is a Process

Leslie M Blaha; Mitchell Abrams; Sarah A Bibyk; Claire Bonial; Beth M Hartzler; Christopher D Hsu; Sangeet Khemlani; Jayde King; Robert St Amant; J Gregory Trafton; Rachel Wong

doi:10.3389/fnsys.2022.800280

Understanding Is a Process

Front Syst Neurosci. 2022 Mar 31:16:800280. doi: 10.3389/fnsys.2022.800280. eCollection 2022.

Authors

Affiliations

¹ 711th Human Performance Wing, U.S. Air Force Research Laboratory, Wright-Patterson Air Force Base, OH, United States.
² Tufts University, Medford, MA, United States.
³ U.S. Army Combat Capabilities Development Command, Army Research Laboratory, Adelphi, MD, United States.
⁴ Link Training & Simulation, CAE USA, Arlington, TX, United States.
⁵ Navy Center for Applied Research in AI, U.S. Naval Research Laboratory, Washington, DC, United States.

Abstract

How do we gauge understanding? Tests of understanding, such as Turing's imitation game, are numerous; yet, attempts to achieve a state of understanding are not satisfactory assessments. Intelligent agents designed to pass one test of understanding often fall short of others. Rather than approaching understanding as a system state, in this paper, we argue that understanding is a process that changes over time and experience. The only window into the process is through the lens of natural language. Usefully, failures of understanding reveal breakdowns in the process. We propose a set of natural language-based probes that can be used to map the degree of understanding a human or intelligent system has achieved through combinations of successes and failures.

Keywords: behavioral measurement; common ground; explainable AI; human-machine teaming; human-robot interaction; mental models; mutual understanding; natural language processing.