SPEECH OGLE: Indexing Uncertainty for Spoken Document Search

Ciprian Chelba, Alex Acero

Abstract: The paper presents the Position Specific Posterior Lattice (PSPL), a novel lossy representation of automatic speech recognition lattices that naturally lends itself to efficient indexing and subsequent relevance ranking of spoken documents.In experiments performed on a collection of lecture recordings --- MIT iCampus data --- the spoken document ranking accuracy was improved by 20% relative over the commonly used baseline of indexing the 1-best output from an automatic speech recognizer.The inverted index built from PSPL lattices is compact --- about 20% of the size of 3-gram ASR lattices and 3% of the size of the uncompressed speech --- and it allows for extremely fast retrieval. Furthermore, little degradation in performance is observed when pruning PSPL lattices, resulting in even smaller indexes --- 5% of the size of 3-gram ASR lattices.