Controlled Language and Baby Turing Test for General Conversational Intelligence

Draft — This content has not been approved for publication.
Authors: Anton Kolonin

← Back to Publications

Controlled Language and Baby Turing Test for General Conversational Intelligence

Authors:


Year: 2020
Venue: arXiv:2005.09280
Links: arXiv abstract

Summary

Proposes a “Baby Turing Test” that extends the classic Turing Test to measure conversational intelligence, paired with a controlled language built on an extensible semantic-graph representation. The author argues that combining a measurable quality test with a controllable, human-inspectable language model offers a path to general-purpose conversational systems that avoid both the brittleness of hand-built dialog trees and the opacity of pure neural chat models.

Relevance to Hyperon

A SingularityNET/Aigents source on interpretable, semantic-graph-based conversational intelligence and controlled-language/interlingua design — relevant to the interpretable-NLP and NL-to-symbolic-structure themes of the Semantic Parsing Deep Dive.

Key References



Discussion