Adam Pease, Richard Thompson
Human language is often vague and ambiguous. There have been many efforts to create formal languages and many attempts to translate human language into formal languages. Logic has a great deal of flexibility, not least in how the symbols used are defined. We anchor lexical elements in a formal ontology, which helps ensure that the meaning of symbols is stable across discourse. We generate a large corpus of language and logic pairs that are used to train a machine learning system. The training data is generated algorithmically from our ontology. Since even a very large training corpus cannot capture all the possible sentences in human language, we employ large language models to simplify language and make it more likely that input sentences will be similar to forms that are in the training data.