The MASTAR Project is a global hub for the speech and text research, with researchers in the field of language processing, machine translation and speech processing from Japan and abroad. It pursues research and development spanning the areas of multilingual speech-to-speech translation, multilingual text translation, multilingual dialog system technologies, and multilingual language resources, within a framework of collaboration between industry, academia and government. The MASTAR Project’s goal is to collect and accumulate speech and text resources, a task that has proved difficult for individual organizations. The MASTAR Project will also accelerate the dissemination of research results back into society for practical application.

MASTAR Project Aims

1.	To establish a global hub for speech/text resources and technologies.
2.	To initiate a sustainable system that stores and develops speech and text resources linking industry and society.
3.	Specifically…
	(1) to carry out research and development and field tests, and to disseminate network-based speech-to-speech translation, one part of the Social Benefit Acceleration Projects of CSTP, Council for Science and Technology Policy. (2) to provide WEB 2.0 machine translation services for areas such as industry sectors and manuals, and to launch a positive development cycle for the accumulation of sharable dictionaries and corpuses for researching translation technologies. (3) to carry out research and development, field tests and dissemination of spoken dialog interface technologies in terms of universal communication to provide information for all users. (4) to create and disseminate global language resources.
4.	To set up an open R&D structure to promote research in collaboration with industry, academia and government, and to develop the human resources to support information technologies in Japan.

---

---

MASTAR Project Director: Dr. Satoshi Nakamura The MASTAR Project is a global hub for the speech and text research, with researchers in the field of language processing, machine translation and speech processing from Japan and abroad. It pursues research and development spanning the areas of multilingual speech-to-speech translation, multilingual text translation, multilingual dialog system technologies, and multilingual language resources, within a framework of collaboration between industry, academia and government. The MASTAR Project’s goal is to collect and accumulate speech and text resources, a task that has proved difficult for individual organizations. The MASTAR Project will also accelerate the dissemination of research results back into society for practical application.

---

Spoken Language Communication Group Group Leader: Dr. Hisashi Kawai The goal of the Spoken Language Communication Group is to achieve natural communication regardless of who or where speakers are, when, how or in which language they speak, with the support of paralinguistic information such as intonation, facial expressions, and gestures.

---

Language Translation Group Group Leader: Dr. Eiichiro Sumita While the language barrier between mother tongue and foreign language is a major challenge in the real world, the automatic translation is expected to play a key role to overcome this barrier. The Language Translation Group is developing a high-quality multi-lingual automatic translation technology, focusing a new core technology called Corpus-based Translation.

---

Language Infrastructure Group Group Leader: Dr. Kentaro Torisawa Language resources refer to a computer compatible dictionary or large amount of text with grammatical and semantic annotation. The Language Infrastructure Group develops, releases and disseminates the language resources and their application tools to contribute to the development of natural language processing technology including machine translation, dialog system, information retrieval and information extraction.