Voice Search
Ye‐Yi Wang, Dong Yu, Yun‐Cheng Ju, Alex Acero · 2011
This chapter reviews the problem of voice search, which is one of the most actively investigated technologies underlying many practical applications. It compares voice search with other spoken language understanding (SLU) technologies for human/computer interaction, discusses challenges in developing voice search applications, and reviews important research work targeting at these problems. As in other SLU applications, robustness is the central issue in voice search. The technology in acoustic modeling aims at improved the robustness to environment noise, different channel conditions and speaker variance; the pronunciation research addresses the problem of unseen word pronunciation and pronunciation variance; the language model research focuses on linguistic variance; the studies in SLU/search give rise to the improved robustness to linguistic variance and automatic speech recognizer (ASR) errors; the dialogue management research enables recovery from the confusions and understanding errors; and the learning in feedback loop speeds up the system tuning for more robust performance. Controlled Vocabulary Terms acoustic signal processing; human computer interaction; speech recognition; voice communication