Results 221 to 230 of about 31,508 (267)

Corrected tandem features for acoustic model training

2008 IEEE International Conference on Acoustics, Speech and Signal Processing, 2008
This paper describes a simple method for significantly improving tandem features used to train acoustic models for large-vocabulary speech recognition. The linear activations at the outputs of an MLP classifier were modified according to known reference labels: where necessary, the activation of the output unit corresponding to the correct phone label ...
Arlo Faria, Nelson Morgan
openaire   +1 more source

Perspectives on Correct Approaches to Training

2007
The life of endurance athletes is about effort, effort during competition and especially effort during training. In preparation for an Olympic competition in the marathon, a race of about 2 hours, an elite runner may train 20 hours per week for nine months or more. In other sports such as cycling, swimming or Nordic skiing, the number of training hours
Carl Foster, Jack T. Daniels, Stephen Seiler   +2 more
openaire   +1 more source

Physical Training in the Correction of Scoliosis

American Physical Education Review, 1913
(1913). Physical Training in the Correction of Scoliosis. American Physical Education Review: Vol. 18, No. 1, pp. 5-5.
openaire   +1 more source

Corrective training of hidden control neural network

Proceedings of ICNN'95 - International Conference on Neural Networks, 2002
A corrective training algorithm for hidden control neural network (HCNN) is proposed in this paper with application to the isolated spoken Korean digit recognition. The proposed algorithm tries to heuristically minimize the number of recognition errors, which improves the discriminatory power of the conventional HCNN-based speech recognizers ...
KyungMin Na, Soo-Ik Chae, SouGuil Ann
openaire   +1 more source

Correction to "On Testing Trained Vector Quantizer Codebooks"

IEEE Transactions on Image Processing, 1997
This correspondence formally defines partial Radon trans- forms for functions of more than two dimensions. It shows that a generalized projection-slice theorem exists which connects planar and hy- perplanar projections of a function to its Fourier transform.
Dong Sik Kim, Taejeong Kim, Sang Uk Lee
openaire   +2 more sources

Improved noise robustness by corrective and rival training

2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221), 2002
We show that discriminative training methods have the potential to improve noise robustness even for high resolution acoustic models trained on noisy data. To this end, we compare the performance of acoustic models trained on noisy data using maximum likelihood (ML), corrective (CT) and rival training (RT).
Carsten Meyer, Georg Rose
openaire   +1 more source

Manpower and Training for Corrections

The Journal of Criminal Law, Criminology, and Police Science, 1968
William E. Amos, Charles Prigmore
openaire   +1 more source

Corrective training for speaker adaptation

6th European Conference on Speech Communication and Technology, 1999
Xiuyang Yu, Wayne H. Ward
openaire   +1 more source

An integrated approach to correctional foodservice training

Journal of the American Dietetic Association, 1985
R D, Reid, C A, Raab, A A, Hertzler, N G, Marriott   +3 more
openaire   +2 more sources

Social Competence Training as a Correctional Alternative

Offender Rehabilitation, 1976
H. R. Novotny, J. J. Enomoto
openaire   +1 more source