Evolution of the performance of automatic speech recognition algorithms in transcribing conversational telephone speech

Conference Record - IEEE Instrumentation and Measurement Technology Conference

Volumn 3, Issue , 2001, Pages 1926-1931

  Padmanabhan, M ^a   Saon, G ^a   Zweig, G ^a   Huang, J ^a   Kingsbury, B ^a   Mangu, L ^a  

^a IBM T J WATSON RESEARCH CENTER   (United States)

Author keywords

Boosting; Consensus; Discriminant transforms; Formant frequencies; Spectral peaks; Speech recognition; Spontaneous speech; Telephone speech

Indexed keywords

ACOUSTICS; DATABASE SYSTEMS; ERROR ANALYSIS; FEATURE EXTRACTION; SPEECH SYNTHESIS; TELEPHONE SWITCHBOARDS;

SPEECH TO TEXT CONVERSION; VOICEMAIL; WORD ERROR RATE;

SPEECH RECOGNITION;

EID: 0034829237     PISSN: None     EISSN: None     Source Type: Conference Proceeding    
DOI: None     Document Type: Conference Paper

References (15)

1. Performance of the IBM large vocabulary speech recognition system on the ARPA wall street journal task
2. Data-driven approach to designing compound words for continuous speech recognition
3. Speaking mode dependent pronunciation modeling in large vocabulary conversational speech recognition
4. Maximum likelihood discriminant feature spaces
5. Use of spectral peak information use in speech recognition
6. Maximum likelihood modeling with gaussian distributions for classification
7. Spectral subband centroid features for speech recognition
8. A course in phonetics
9. Resonator-based filter-banks for frequency domain applications
10. Boosting gaussian mixtures in an LVCSR system
11. Boosting the margin: A new explanation for rge effectiveness of voting methods
12. Finding consensus among words: Lattice-based word error minimization
13. A post-processing system to yield reduced word error rates: Recogniser output voting error reduction (ROVER)
14. Performance improvement in voicemail transcription