ATR has built a multi-language speech translation system called ATR-MATRIX. It consists of a spoken-language translation subsystem, which is the focus of this paper, together with a highly accurate speech recognition subsystem and a high-definition speech synthesis subsystem. This paper gives a road map of solutions to the problems inherent in spoken-language translation. Spoken-language translation systems need to tackle difficult problems such as ungrammaticality. contextual phenomena, speech recognition errors, and the high-speeds required for real-time use. We have made great strides towards solving these problems in recent years. Our approach mainly uses an example-based translation model called TDMT. We have added the use of extra-linguistic information, a decision tree learning mechanism, and methods dealing with recognition errors.

Recently, several types of Japanese to English MT (machine translation) systems have been developed, but prior to using such systems, they have required a pre-editing process of re-writing the original text into Japanese that could be easily translated. For communication of translated information requiring speed in dissemination, application of these systems would necessarily pose problems. To overcome such problems, a Multi-Level Translation Method based on Constructive Process Theory had been proposed. In this paper, the benefits of this method in ALT-J/E will be described. In comparison with the conventional elementary composition method, the Multi-Level Translation Method, emphasizing the importance of the meaning contained in expression structures, has been ascertained to be capable of conducting translation according to meaning and context processing with comparative ease. We are now hopeful of realizing machine translation omitting the process of pre-editing.