Towards Automated Translating Natural Language Sentences into Intensional Constructions

Autor: Branislav Bednar, Zuzana Bilanova
Rok vydání: 2019
Předmět:
Zdroj: 2019 IEEE 15th International Scientific Conference on Informatics.
DOI: 10.1109/informatics47936.2019.9119252
Popis: This paper aims to introduce a new and original implementation of the semantic machine of transparent intensional logic. This logic system is characterized by its procedural semantics, thanks to which it is used in the field of logical analysis of natural language. Transparent intensional logic is a high-order logic, where partial functions of the intensional typed λ-calculus allow to remove the semantic ambiguities of natural language. Natural language expressions are usually analyzed in three steps - type analysis, synthesis of construction, and type control, which is an optional step. The presented implementation of the semantic machine of transparent intensional logic is focused on the first two of the mentioned steps of formalizing the meaning of sentences. This semantic machine uses outputs from the Stanford CoreNLP parser. Its functionality and correctness are demonstrated at the end of the paper through the analysis of several sentences of natural language.
Databáze: OpenAIRE