Software Development Guidelines

{ Japanese / English }

The "Basic Software for Anthropomorphic Spoken Dialogue Agents", developed in the project of IPA (Information-Technology Promotion Agency, Japan) known as Galatea Project, consists of modules of speech recognition, speech synthesis, facial image synthesis, and dialogue integration. At the Consortium, besides providing as open software a basic software toolkit indispensable to interactive speech technologies, in order to improve and maintain it continuously, and to make it spread as the baseline/reference system of interactive speech, further expansion is promoted from the following viewpoints.
  1. Maintenance of the common base of fundamental research
    Interactive speech technologies make necessary many dialogue modules. Besides, in order to realize a natural dialogue, not only the function and performance of the individual modules, but also advanced dialogue administration ability and a dialogue description means are indispensable. The Consortium is working on the maintenance and expansion of the spoken dialogue reference system that can become the base of research and development.

  2. Improvement of the function and performance of dialogue modules
    In order that they can become as a whole the baseline/reference software of research and development, each of the software modules of speech recognition, speech synthesis, facial image synthesis, and dialogue integration are working on an even higher functionality and efficiency. Moreover, the efforts are concentrating on their tuning as dialogue system components necessary in terms of realizing a natural dialogue in a real environment.

  3. Improvement of the convenience for other fields, industry, etc.
    To support the construction of spoken dialogue applied systems that use interactive speech technologies (IVR: Interactive Voice Response), as well as multimodal dialogue applied systems (MMI: Multi-modal Interaction), the following improvements are being performed:

    1. Transplantation of individual modules to Windows
    2. Correspondence to the standard API
    3. Expansion of the VoiceXML standard base
    4. Improvement of development environment (prototyping tools, etc.)
    5. Completion of documentation
Return to ISTC top page