Speech and language technologies for audio indexing and retrieval

Makhoul, John; Kubala, Francis; Leek, Tim; Liu, Daben; Nguyen, Long; Schwartz, Richard; Srivastava, Amit

doi:10.1109/5.880087

Cited by 125 publications

(64 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…al., 1996b) to more recent systems based on large scale broadcast news transcription systems working in conjunction with information retrieval systems (Thong et. al., 2000, Makhoul et. al., 2000.…”

Section: Spoken Document Retrievalmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Scale Spoken Document Retrieval for Cantonese Broadcast News

Meng

Ching

2004

International Journal of Speech Technology

View full text Add to dashboard Cite

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.

show abstract

Section: Spoken Document Retrievalmentioning

confidence: 99%

“…al., 1999), SpeechBot of Compaq (Thong et. al., 2000 andCompaq, 2000), Rough'n Ready of BBN (Makhoul et. al., 2000), and Multimedia Document Retrieval project of Cambridge University Tuerk et.…”

Section: Spoken Document Retrievalmentioning

confidence: 99%

Multi-Scale Spoken Document Retrieval for Cantonese Broadcast News

Meng

Ching

2004

International Journal of Speech Technology

View full text Add to dashboard Cite

show abstract

“…Current approaches have significant shortcomings. Most methods are either rule-based [5], or require significant amounts of manually labeled training data to achieve a reasonable level of performance [4]. The methods may identify a name, company, or location, but this is only a small part of the information that should be extracted; we would like to know further, for example, that a particular person is a politician and that a location is a vacation resort.…”

Section: Named Entity Extractionmentioning

confidence: 99%

“…While rule-based systems suffer significant degradations in going from mixed case to this style of text, hidden Markov model (HMM) approaches have proven to be more robust, suffering only a degradation of 4% to 5% in F-measure [4]. We will implement a variation of the HMM approach in which the output distributions are exponential models that weight various features of the words (numbers, titles, etc.…”

Section: Named Entity Extractionmentioning

confidence: 99%

Enhanced perspectives for historical and cultural documentaries using informedia technologies

Wactlar

Chen

2002

Proceedings of the 2nd ACM/IEEE-CS Joint Conference on Digital Libraries

View full text Add to dashboard Cite

“…The real-time audio processing in the eViTAP system is performed by the BBN AudioIndexer system, described in detail in (Makhoul et al 2000). The AudioIndexer system provides a wide range of real-time audio processing components, including automatic speech recognition, speaker segmentation and identification, topic classification, and named entity detection.…”

Section: Real-time Spoken Language Processingmentioning

confidence: 99%