Automatic Collection of Transcribed Speech for Low Resources Languages

Abstract

Speech is a crucial for human communication and combined with the evolution of instant messaging in voice format as well as automated chatbots, its importance is greater. While the majority of speech technologies have achieved high accuracy, they fail when tested for accents that deviate from the "standard" of a language. This becomes more concerning for languages that lack on datasets and have scarce literature, like Brazilian Portuguese. Thus, this paper proposes a methodology to collect and release a speech dataset for Brazilian Portuguese. The method explores the availability of data and information in video platforms, and automatically extracts the audio from TEDx Talks.