Discriminative training of self-structuring hidden control neural models
نویسندگان
چکیده
This paper presents a new training algorithm for Self-structuring Hidden Control Neural (SHC) models, which we presented at ICASSPCl]. The SHC models were trained non-discriminatively for speech recognition applications [2]. Better recognition performance can generally be achieved, if discriminative training is applied in stead. Thus we developed a discriminative training algorithm for SHC models, where each SHC model for a specific speech pattern is trained with utterances of the ]pattern to be recognized and with other utterances. The discriminative training of SHC neural models has been tested on the TIDIGITS database [3].
منابع مشابه
Discriminative Training of a Neural Predictive System for on-line Word Recognition
We first present the general formalism for training a neural predictive system for on-line word recognition. Our approach implements a Hidden Markov Model (HMM)-based cooperation of several multilayer neural networks. Successive parts of a word trajectory are this way modeled by different Neural Networks [6]. A dynamical segmentation allows to adjust neural letter-models to each word. Training ...
متن کاملAdaptive discrimination in an HMM-based neural predictive system for on-line word recognition
In a recent publication [3], we have introduced a neural predictive system for on-line word recognition. Our approach implements a Hidden Markov Model (HMM)-based cooperation of several neural networks. The task of the HMM is to guide the training procedure of neural networks on successive parts of a word. Each word is modeled by the concatenation of letter-models corresponding to the letters c...
متن کاملRelative Density Nets: A New Way to Combine Backpropagation with HMM's
Geoffrey E. Hinton Gatsby Unit, UCL London, UK WCIN 3AR [email protected] Logistic units in the first hidden layer of a feedforward neural network compute the relative probability of a data point under two Gaussians. This leads us to consider substituting other density models. We present an architecture for performing discriminative learning of Hidden Markov Models using a network of many...
متن کاملGenerative Kernels and Score-Spaces for Classication of Speech: Progress Report iii
May is is the third and nal progress report for Project /// (Generative Kernels and Score Spaces for Classiication of Speech) within the Global Uncertainties Programme. is project combines the current generative models developed in the speech community with discriminative classiiers. An important aspect of the approach is that the generative models are used to deene a score-space that can be us...
متن کاملGMM-Free Flat Start Sequence-Discriminative DNN Training
Recently, attempts have been made to remove Gaussian mixture models (GMM) from the training process of deep neural network-based hidden Markov models (HMM/DNN). For the GMM-free training of a HMM/DNN hybrid we have to solve two problems, namely the initial alignment of the frame-level state labels and the creation of context-dependent states. Although flat-start training via iteratively realign...
متن کامل