Commencement of joint research on voice-cloning from a handful of human voice samples with Associate Professor Junichi Yamagishi of the National Institute of Informatics (NII)

Elements expressing users’ personality Voice cloning Progressing with natural language processing research and developments

Alt, Inc. which has developed ”al+”, the Personal Artificial Intelligence (P.A.I.) (Headquarters: Koto-ku, Tokyo, CEO, Mr. Kazutaka Yonekura) has begun a joint research on voice-cloning with Associate Professor Junichi Yamagishi of the National Institute of Informatics (hereafter known as “NII”).

Alt believes that in the development of P.A.I., it is equally essential to study about outer appearance of human beings as the thinking and communication function. The human voice is a very important element to establish the identity of its owner.

The voice synthesis technology which Yamagishi is studying reproduces unique audio characteristics of a person’s voice by measuring the difference from derived mean value of a large human voice sample collection.

While a vast amount of voice recordings is required in the conventional synthetic speech generation in order to train the computer to learn the unique characteristics of a speaker’s voice, high precision voice-cloning is made possible using a small amount of voice recordings with the application of Yamagishi's technology

In this joint research, we will be conducting research and development to integrate the P.A.I. which Alt is developing and Yamagishi’s voice-cloning technology, to create a system which automatically generates advanced cloned avatars. Further, the said acoustic model will be applied to a large number of users simultaneously, to build a system which can generate a large amount of voice clones in a short period of time.


About Yamagishi Junichi


Research Organization of Information and systems(ROIS) National Institute of Informatics Associate Professor, Digital Content and Media Sciences Research Division
Degree:Doctoral degree(Engineering)
Research Fields: Speech information processing / Speech synthesis / Speech-based human machine interaction

Completed PhD degree in engineering, Tokyo Institute of Technology, Japan, 2006. Visiting researcher at University of Edinburgh, 2006 ~ 2007. Research Fellow at University of Edinburgh, 2007 ~ 2013. Associate Professor at Digital Content and Media Sciences Research Division, National Institute of Informatics cum Senior Research Fellow at the Centre for Speech Technology Research, University of Edinburgh.

Conduct research in speech information processing. Competed in Japanese-style statistical speech synthesis in an effort to spread its influence globally through European Union’s academic projects. Further, developed a VoiceBank project which is a speech synthesis’ welfare application technology in United Kingdom and Japan. Won the Itakura Prize (Innovative Young Researcher Award) from the Acoustical Society of Japan for the achievements in adaptive speech synthesis in 2010 and The Young Scientists’ Prize for the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology in 2014.