Natural Language Processing (NLP) is one of the branches of Artificial Intelligence focused on the analysis of human language. It is divided into different levels of processing: Phonology, Morphology and Lexical.
The first, Phonology, is about the interpretation of sound. It considers phonetic and phonemic rules and prosody (the intonation and rhythm of speech). Morphology deals with morphemes, the small units of meaning of a word, for example the understanding of prefixes and suffixes. Lexical language processing interprets the meanings of the word itself, taking into account the context of application. Interpretation at the lexical level requires syntactic, semantic, discursive, and pragmatic knowledge.
In voice assistants, NLP works in three steps: First, it converts speech to text. The system processes the set of sounds and transforms them into words. After that, the system processes the words and triggers other areas of Artificial Intelligence to search for the answer. Finally, the voice assistant transforms the words into sound to respond to the user.
NLP-based systems need to be trained. For a machine to recognize, “understand”, “interpret”, and answer a human query, it needs to have a large database, as diverse as possible.
