Bottle-Neck Feature Extraction Structures for Multilingual Training and Porting
نویسندگان
چکیده
Stacked-Bottle-Neck (SBN) feature extraction is a crucial part of modern automatic speech recognition (ASR) systems. The SBN network traditionally contains a hidden layer between the BN and output layers. Recently, we have observed that an SBN architecture without this hidden layer (i.e. direct BN-layer – output-layer connection) performs better for a single language but fails in scenarios where a network pre-trained in multilingual fashion is ported to a target language. In this paper, we describe two strategies allowing the direct-connection SBN network to indeed benefit from pre-training with a multilingual net: (1) pre-training multilingual net with the hidden layer which is discarded before porting to the target language and (2) using only the the directconnection SBN with triphone targets both in multilingual pre-training and porting to the target language. The results are reported on IARPA-BABEL limited language pack (LLP) data. c © 2016 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the Organizing Committee of SLTU 2016.
منابع مشابه
Multilingual Deep Bottle Neck Features a Study on Language Selection and Training Techniques
Previous work has shown that training the neural networks for bottle neck feature extraction in a multilingual way can lead to improvements in word error rate and average term weighted value in a telephone key word search task. In this work we conduct a systematic study on a) which multilingual training strategy to employ, b) the effect of language selection and amount of multilingual training ...
متن کاملImproving Deliverable Speech-to-Text Systems with Multilingual Knowledge Transfer
This paper reports our recent progress on using multilingual data for improving speech-to-text (STT) systems that can be easily delivered. We continued the work BBN conducted on the use of multilingual data for improving Babel evaluation systems, but focused on training time-delay neural network (TDNN) based chain models. As done for the Babel evaluations, we used multilingual data in two ways:...
متن کاملHierarchical bottle neck features for LVCSR
This paper investigates the combination of different neural network topologies for probabilistic feature extraction. On one hand, a five-layer neural network used in bottle neck feature extraction allows to obtain arbitrary feature size without dimensionality reduction by transform, independently of the training targets. On the other hand, a hierarchical processing technique is effective and ro...
متن کامل(Deep) Neural Networks
This work continues in development of the recently proposed Bottle-Neck features for ASR. A five-layers MLP used in bottleneck feature extraction allows to obtai arbitrary feature size without dimensionality reduction by transforms, independently on the MLP training targets. The MLP topology – number and sizes of layers, suitable training targets, the impact of output feature transforms, the ne...
متن کاملAn Investigation on Initialization Schemes for Multilayer Perceptron Training Using Multilingual Data and Their Effect on ASR Performance
In this paper we present our latest investigation on initialization schemes for Multilayer Perceptron (MLP) training using multilingual data. We show that the overall performance of a MLP network improves significantly by initializing it with a multilingual MLP. We propose a new strategy called “open target language” MLP to train more flexible models for language adaptation, which is particular...
متن کامل