Bypassing words in automatic speech recognition

Automatic speech recognition (ASR) is usually defined as the transformation of an acoustic signal to words. Though there are cases where the transformation to words is useful, the definition does not exhaust all contexts in which ASR could be used. Once the constraint that an ASR system outputs words is relaxed, modifications that reduce the search space become possible: 1) The use of syllables instead of words in the recognizer's language model; 2) The addition of a concept model that transforms syllable strings to concept strings, where a concept collects related words and phrases. The paper presents preliminary positive results on the use of syllables and concepts in speech recognition and outlines our current efforts to verify the Syllable-Concept Hypothesis (SCH).