Read more
Traditionally, the European-based biannual international conference "EUROSPEECH" dealing with all aspects of speech science and technology is preceded by an "ESPRIT Speech Projects Days", which presents a particularly well timed opportunity to measure progress in speech technology and ap plications in Europe. The last venue was held in Berlin, Germany, on September 20th, 1993. The success of this workshop encouraged the major European experts in the field to contribute to this volume. Published in the ESPRIT Research Report series, it presents the results of advanced European research on speech technologies and its applications in the multilingual framework of the European Union. Speech is an important factor in building an integrated European communication platform. Strong links exist between speech and natural language processing, and human computer interaction. Recent experimental results on multilingual conversion between both speech and text show the advantage of integrating phonetic, lexical, and syntactic knowledge, and also demonstrate the feasibility of multilingual voice systems in the human-computer interface applications. Multilingual queries use natural language-based co-operative dialogue as an interface to the computer services in the information applications. Continuous and robust speech understanding is here addressed for both speaker-independent and speaker-adaptive processing, together with dialogue modelling and manage ment. Such technologies are then used in the design of computer workstations with a speech-based human interface for a large range variety of information technology applications (e.g. in the office, telecommunications, and computer aided education).
List of contents
Advanced Speech Technologies in the ESPRIT Programme.- A Common European Approach to Assessment, Corpora and Standards.- Robust System to Access Large Databases in Natural Language.- SPELL: A Pronunciation Training Device Based on Speech Technology.- A Multilingual Unlimited Vocabulary Stochastic Tagger.- SYMBOL: Multilingual and Multicode Lexical Learning System on CD-I Environment.- Speech Understanding and Dialogue over the Telephone.- Dialogue Development and Implementation in the Danish Dialogue Project.- SUNSTAR.- The EAGLES Working Group on Spoken Language.- Recent Progress in Speech-to-Text Conversion at LIMSI.- Sound-to-Gesture Inversion in Speech: The Speech Maps Approach.- ESPRIT II Project 5516 ROARS: Robust Analytic Speech Recognition System.- POLYGLOT Project: Hybrid System NN/HMM for Continuous Speech Recognition.- Principles and Applications of the VINICS Continuous-Speech Recognition System.- The Philips Research System for Large-Vocabulary and Continuous-Speech Recognition.- DANDELION: Variations of Discourse Functions and Representations According to Context.- A Friendly Interactive Robot for Service Tasks.- Architectural Design of the PLUS Dialogue System.- WERRNIKE: A Neural Network-Based, Speaker-Independent, Large-Vocabulary, Continuous-Speech Recognition System.- Contributors.
