Language understanding as a step towards human level intelligence - automatizing the construction of the initial dictionary from example sentences

TL;DR

This paper presents a novel approach to automatically construct initial dictionaries for natural language understanding systems by analyzing sentence-meaning pairs, improving the translation of natural language into formal representations.

Contribution

It introduces an improved method for learning initial lexicons from training data, enhancing the process of translating natural language into formal language.

Findings

The method outperforms existing approaches on database querying tasks.

It effectively learns initial lexicons from sentence-meaning pairs.

The approach improves natural language understanding for robot commands.

Abstract

For a system to understand natural language, it needs to be able to take natural language text and answer questions given in natural language with respect to that text; it also needs to be able to follow instructions given in natural language. To achieve this, a system must be able to process natural language and be able to capture the knowledge within that text. Thus it needs to be able to translate natural language text into a formal language. We discuss our approach to do this, where the translation is achieved by composing the meaning of words in a sentence. Our initial approach uses an inverse lambda method that we developed (and other methods) to learn meaning of words from meaning of sentences and an initial lexicon. We then present an improved method where the initial lexicon is also learned by analyzing the training sentence and meaning pairs. We evaluate our methods and compare them with other existing methods on a corpora of database querying and robot command and control.