ここまで来た音声技術・今後の展望 / Current progress on speech technologies and its future prospects @ LINE DEV DAY 2020

Abstract

Speech processing technologies such as automatic speech recognition (ASR) and speech synthesis has rapidly progressed. Furthermore, new research fields such as acoustic scene/event detection have appeared. In this session, two engineers from LINE (Yusuke Kida and Ryuichi Yamamoto) will talk about current progress of ASR and speech synthesis. From Doshisha University, associate prof. Keisuke Imoto will talk about current progress of acoustic scene/event detection in which Japanese researchers made remarkable contributions, e.g., invitation of a flagship conference DCASE. Although progress of deep learning has a strong influence, we would like to deep dive into how specific elements in speech processing connects with deep learning and leads to technological progress. We would also like to discuss about common technological elements among speech processing fields and specific elements to clarify characteristics of each field, and we will discuss about future prospects.