Use of non-negative matrix factorization for language model adaptation in a lecture transcription task

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

Volumn 1, Issue , 2001, Pages 541-544

  Novak, Miroslav ^a   Mammone, Richard ^b  

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

^b RUTGERS UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

LANGUAGE MODEL ADAPTATION; NON-NEGATIVE MATRIX FACTORIZATION; SINGULAR VALUE DECOMPOSITION;

COMPUTATIONAL LINGUISTICS; DATABASE SYSTEMS; MATRIX ALGEBRA; PROBLEM SOLVING; VECTORS;

SPEECH RECOGNITION;

EID: 0034854666     PISSN: 15206149     EISSN: None     Source Type: Journal    
DOI: 10.1109/ICASSP.2001.940887     Document Type: Article

References (7)

1. A multispan language modelling framework for large vocabulary speech recognition
2. Learning the parts of objects by non-negative matrix factorization
3. Trigger-based language models using maximum likelihood estimation of exponential distributions
4. Putting it all together: Language model combination
5. The Aristotle project: A distributed learning system, proceedings of Ed-Media2000
6. Statistical machine translation as non-negative matrix factorization