Name
Abstract | ||
|---|---|---|
We have recently extended the capabilities of BBN's large vocabulary discrete-utterance speech recognition system (BYBLOS) to operate on raw audio recordings of radio news programming. The recording are given to the system as large monolithic waveforms without any additional side-information. Our goal is to transcribe ail speech in the input with the highest accuracy possible. The problem is very challenging because radio news programming has frequent changes in speaker, speaking style, dialect, accent, topic, channel, and environmental conditions. Furthermore, the monolithic input presents new problems for recognition algorithms and language models since all useful boundaries (such as speaker turns or sentence ends) are unknown. |
Year | DOI | Venue |
|---|---|---|
1996 | 10.1109/ICSLP.1996.607432 | ICSLP 96 - FOURTH INTERNATIONAL CONFERENCE ON SPOKEN LANGUAGE PROCESSING, PROCEEDINGS, VOLS 1-4 |
Keywords | Field | DocType |
hidden markov models,channel,speech recognition,language model,system testing,accuracy,background noise,training data,telephony,natural languages,bandwidth,audio recording,radio broadcasting,language models | Radio broadcasting,Transcription (linguistics),Computer science,Audio mining,Speech recognition,Raw audio format,Speaker recognition,Artificial intelligence,Natural language processing,Sentence,Language model,Acoustic model | Conference |
Citations | PageRank | References |
2 | 0.90 | 3 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Francis Kubala | 1 | 410 | 99.88 |
| Tasos Anastasakos | 2 | 412 | 56.01 |
| Hubert Jin | 3 | 36 | 7.12 |
| Long Nguyen | 4 | 326 | 84.60 |
| Richard M. Schwartz | 5 | 2839 | 717.76 |