Multi-task learning strategies for a recurrent neural net in a hybrid tied-posteriors acoustic model

9th European Conference on Speech Communication and Technology

Volumn , Issue , 2005, Pages 2993-2996

  Stadermann, Jan ^a   Koska, Wolfram ^a   Rigoll, Gerhard ^a  

^a TECHNICAL UNIVERSITY OF MUNICH   (Germany)

Author keywords

[No Author keywords available]

Indexed keywords

ACOUSTIC PROPERTIES; MARKOV PROCESSES; MATHEMATICAL MODELS; NEURAL NETWORKS; PERSONNEL TRAINING; PROBABILITY; SPEECH RECOGNITION;

AUTOMATIC CLASSIFIERS; HMM STATE POSTERIOR PROBABILITIES; HYBRID ACOUSTIC MODELS; MULTI-TASK LEARNING STRATEGIES;

LEARNING ALGORITHMS;

EID: 33745216217     PISSN: None     EISSN: None     Source Type: Conference Proceeding    
DOI: None     Document Type: Conference Paper

References (10)

1. Herve Bourlard and Nelson Morgan, Connectionist Speech Recognition: A Hybrid Approach, Kluwer Academic Publishers, 1994.
2. M. Hochberg, G. Cook, S. Renals, A. Robinson, and R. Schechtman, "ABBOT Hybrid Connectionist-HMM Large-Vocabulary Recognition System," in Spoken Language Systems Technology Workshop, 1995, pp. 170-176.
3. Jörg Rottland and Gerhard Rigoll, "Tied Posteriors: An Approach for Effective Introduction of Context Dependency in Hybrid NN/HMM LVCSR," in IEEE Int. Conference on Acoustics, Speech, and Signal Processing (ICASSP), Istanbul, Turkey, June 2000.
4. Jan Stadermann and Gerhard Rigoll, "Comparing NN Paradigms in Hybrid NN/HMM Speech Recognition using Tied Posteriors," in IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU), St. Thomas, U.S. Virgin Islands, Nov. 2003.
5. Shahla Parveen and Phil Green, "Multitask Learning in Connectionist Robust ASR using Recurrent Neural Networks," in European Conference on Speech Communication and Technology, Geneva, Switzerland, Sept. 2003, pp. 1813-1816.
6. A. J. Robinson, "An Application of Recurrent Nets to Phone Probability Estimation," IEEE Transactions on Neural Networks, vol. 5, no. 2, pp. 298-305, Mar. 1994.
7. S. Santini and A. Del Bimbo, "Recurrent neural networks can be trained to be maximum A posteriori probability classifiers," Neural Networks, vol. 8, no. 1, pp. 25-29, 1995.
8. Christian Igel and Michael Hüsken, "Improving the Rprop Learning Algorithm," in Proceedings of the Second International Symposium on Neural Computation NC'2000. 2000, pp. 115-121, ICSC Academic Press.
9. Rich Caruana, "Multitask Learning," Machine Learning, no. 28, pp. 41-75, 1997.
10. Mirjam Killer, Sebastian Stüker, and Tanja Schultz, "Grapheme Based Speech Recognition," in European Conference on Speech Communication and Technology, Geneva, Switzerland, Sept. 2003, pp. 3141-3144.