Natural language understanding (NLU, not NLP) in cognitive systems

Abstract

Developing cognitive agents with human-level natural language understanding (NLU) capabilities requires modeling human cognition because natural, unedited utterances are anything but neat and complete; so understanding them requires the ability to clean up the input and fill in the lacunae. Indeed, language inputs regularly feature complex linguistic phenomena such as lexical and referential ambiguity, ellipsis, false starts, spurious repetitions, semantically vacuous fillers, nonliteral language, indirect speech acts, implicatures, and production errors. Moreover, cognitive agents must be nimble in the face of incomplete interpretations since even people do not perfectly understand every aspect of every utterance they hear. This means that once an agent has reached the best interpretation it can, it must determine how to proceed - be that acting upon the new information directly, remembering whatever it has understood and waiting to see what happens next, seeking out information to fill in the blanks, or asking its interlocutor for clarification. The reasoning needed to support NLU extends far beyond language itself, including, nonexhaustively, the agent's understanding of its own plans and goals; its dynamic modeling of its interlocutor's knowledge, plans, and goals, all guided by a theory of mind; its recognition of diverse aspects of human behavior, such as affect, cooperative behavior, and the effects of cognitive biases; and its integration of linguistic interpretations with its interpretations of other perceptive inputs, such as simulated vision and nonlinguistic audition. Considering all of these needs, it seems hardly possible that fundamental NLU will ever be achieved through the kinds of knowledge-lean text-string manipulation being pursued by the mainstream natural language processing (NLP) community. Instead, it requires a holistic approach to cognitive modeling of the type we are pursuing in a paradigm called OntoAgent.