Deep Learning Based Speech Recognition: A Review Paper

Abstract

The act of converting audible words into written ones is called speech recognition. To recognize speech, sound waves must be captured and digitally transformed into basic digital representations. Examining the contexts and phonemes of words to ensure the correct spelling for words with comparable sounds is key when creating words from linguistic components or phonemes. The purpose of the paper is to review pattern matching and deep learning-based speech recognition methods explored in recent years. Intelligent virtual assistants or intelligent personal assistants are terms used to describe software agents that carry out tasks or offer services on behalf of users. Private supported commands or queries are very essential in modern speech recognition systems. When the phrase is exclusively accessed through online chat or asking generic questions of virtual assistants, then the online chat programs can occasionally simply be used for fun.