Stochastic Pronunciation Modeling for Out-of-Vocabulary Spoken Term Detection

Abstract

Abstract-Spoken term detection (STD) is the name given to the task of searching large amounts of audio for occurrences of spoken terms, which are typically single words or short phrases. One reason that STD is a hard task is that search terms tend to contain a disproportionate number of out-of-vocabulary (OOV) words. The most common approach to STD uses subword units. This, in conjunction with some method for predicting pronunciations of OOVs from their written form, enables the detection of OOV terms but performance is considerably worse than for in-vocabulary terms. This performance differential can be largely attributed to the special properties of OOVs.One such property is the high degree of uncertainty in the pronunciation of OOVs. We present a stochastic pronunciation model (SPM) which explicitly deals with this uncertainty. The key insight is to search for all possible pronunciations when detecting an OOV term, explicitly capturing the uncertainty in pronunciation. This requires a probabilistic model of pronunciation, able to estimate a distribution over all possible pronunciations. We use a joint-multigram model (JMM) for this and compare the JMM-based SPM with the conventional soft match approach. Experiments using speech from the meetings domain demonstrate that the SPM performs better than soft match in most operating regions, especially at low false alarm probabilities. Furthermore, SPM and soft match are found to be complementary: their combination provides further performance gains.

Handling overlaps in spoken term detection

“…We tested two scenarios for OOV terms: one with 1-best pronunciations and the other with multiple pronunciations based on SPM [9]. We define a maximum group of overlapping detections as a cluster, the averaged number of detection per cluster as the 'overlap ratio', and the size of the largest cluster as 'max overlap'.…”

“…We have shown in previous studies [7] that an ATWV-oriented decision strategy [8] is essential for OOV STD, however accumulated confidences are not normalized and thus cannot be applied together with ATWV-oriented decisions, although normalization techniques may provide some compensation [7]. Furthermore, stochastic pronunciation modeling (SPM) [9], which has been shown to be highly effective for OOV term detection, may further complicate the pattern of overlaps since more pronunciation candidates are taken into account.…”

Spoken Term Detection Using Phoneme Transition Network from Multiple Speech Recognizers' Outputs

“…There have been many studies on speech recognition errors, OOV, and term pronunciation problems in STD [11], [12], [13]. This study addresses only the recognition errors and OOV problems.…”

G2p variant prediction techniques for ASR and STD

“…These variants can occur in a continuum ranging from generally accepted alternate pronunciations to barely perceptible phonetic variations. While incorporating necessary variants can improve both automatic speech recognition (ASR) and spoken term detection (STD) performance [1,2], introducing superfluous variants can lead to increased confusability and a decrease in performance [3]. The use of pronunciation variants in state-of-the-art speech-recognition systems has received significant attention in recent years, with a number of studies investigating the effect of variant prediction on either ASR or STD accuracy [4,5,1,3,6].…”