We focus on speech, which is the most natural and efficient means of communication for humans, and conduct research on its analysis, understanding, and modeling. In particular, we aim to realize more natural and seamless communication between humans, as well as between humans and computers, by modeling the diverse acoustic, linguistic, and affective information contained in human speech through engineering-based approaches. In addition, by scientifically investigating the characteristics of human speech and the mechanisms underlying communication, we seek to address the fundamental question: “What is speech?”. Specifically, we conduct engineering-oriented research to advance core speech processing technologies such as automatic speech recognition, speech synthesis, speech enhancement, and dialogue management. At the same time, we carry out scientific research on the detailed analysis of, e.g., children’s speech and stuttered speech. Furthermore, we apply the outcomes of these studies to a wide range of fields, including education, healthcare, and welfare.