USE OF ASSOCIATIVE INFORMATION TO CONTINUOUS SPEECH RECOGNITION

The continuous speech frequently includes intervals where the utterer is ''lazy'' and does not exhibit a precise pronunciation. Sometimes th e accurate recognition of auxiliaries, which are grammatically importa nt, is prevented. Consequently, to recognize the continuous speech, it is necessary to make use of more intuitive information, i.e., the ass ociative information as the human does, in addition to linguistic info rmation such as syntax and semantics. First, this paper shows that the continuous speech contains ambiguously uttered intervals, for which t he acoustic processing is difficult. Next, the human associations amon g the words are examined and investigated, aiming at the construction of the associative word dictionary to be utilized in the speech recogn ition. Using such an associative word dictionary, the association stre ngth between words is determined, and the succeeding word is predicted . The associative information, not only the acoustic information, also is included in the recognition of the phrase. In the previous speech recognition system by the authors, the phoneme recognition rate of app roximately 90 percent is required in order to achieve the sentence rec ognition rate of 75 percent, if the associative information is not uti lized. The phoneme recognition rate of approximately 80 percent is suf ficient to achieve the sentence recognition rate of 75 percent by util izing the associative information. It is seen also in the system const ructed by the authors that the acoustic information and the associativ e information should be utilized by the ratio of approximately 3:1, fo r scoring the information.