SRI International has developed a spoken-language instruction system that uses audio input and output in foreign language teaching. The system presents information to a learner in either graphical or audio form, and the learner responds by speaking. The learner's speech to the system is recognized and evaluated by a modified version of SRI's DECIPHER HMM-based speech recognition system. The system has been implemented on commercially available microcomputer hardware and responds in real time to the learner's speech. The recognition system was modified for non-native speech. The system environment supports several kinds of instruction in spoken language, including spoken responses to multiple choice questions as well as practice in the production of syntactic patterns and conversational formulas. Several example lessons have been developed for this voice-input environment. The Voice Interactive Language Instruction (VDLJ TM) system itself is demonstrated as part of the presentation. Keywords: language, instruction, speech recognition
Bibliographic reference. Bernstein, Jared / Rtischev, Dimitry (1991): "A voice interactive language instruction system", In EUROSPEECH-1991, 981-984.