Multi-Scale Spoken Document Retrieval for Cantonese Broadcast News

Abstract

This paper presents the application of a multi-scale paradigm to Chinese spoken document retrieval (SDR) for improving retrieval performance. Multi-scale refers to the use of both words and subwords for retrieval. Words are basic units in a language that carry lexical meaning and subword units (such as phonemes, syllables or characters) are building components for words. Retrieval using subword indexing units is found to perform better than words because of the robustness of subword units to out-of-vocabulary (OOV) words during speech recognition and ambiguities in word segmentation. Experimental results have demonstrated that subword bigrams can bring improvement in retrieval performance over words (~9.56%). Application of multi-scale fusion to SDR aims at combining the lexical information of words and the robustness of subwords. This work presents the first detailed investigation for a Cantonese broadcast news retrieval task using two different multi-scale fusion approaches: pre-retrieval fusion and post-retrieval fusion.Multi-scale retrieval using both words and syllable bigrams achieve improvement in retrieval performance (~1.90%) over retrieval on the composite scales.

On the effectiveness of subwords for lexical cohesion based story segmentation of Chinese broadcast news

“…phonemes, syllables and sub-phonetic segments) have shown robustness to speech recognition errors and OOV words in spoken document retrieval (SDR) [23] tasks. Especially for Chinese, retrieval based on character or syllable indexing is superior to words due to the special features of Chinese [5,22]. We believe that subwords should also be effective in story segmentation of erroneous broadcast news transcripts through partial matching.…”

MAHIR System: Unsupervised Segmentation for Malay Spoken Broadcast News Stories

“…In order to process Turkish language documents recognition and indexing units are used as sub-words units by Parlak et al to reduce both the OOV rate and the index alternative recognition hypothesis to handle ASR errors [26]. Some researcher such as Lo et al [4] concentrated on the application of a multi-scale paradigm for Chinese SDR to simply improve retrieval performance. BASRAH [18] system has been designed to detect story boundaries in multilingual (English and Malay) using Confidence Measures (CMs) of the ASR.…”

“…A spoken document retrieval (SDR) system uses automatic speech recognition and information retrieval technologies to analyze and process multimedia documents [1]- [4]. Automatic speech recognition (ASR) systems are used to convert spoken documents (speech) into text transcription.…”