conftrace_

2017 Interspeech INTERSPEECH 2017

Gate Activation Signal Analysis for Gated Recurrent Neural Networks and its Correlation with Phoneme Boundaries

Abstract

In this paper we analyze the gate activation signals inside the gated recurrent neural networks, and find the temporal structure of such signals is highly correlated with the phoneme boundaries. This correlation is further verified by a set of experiments for phoneme segmentation, in which better results compared to standard approaches were obtained.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Speech & Audio

Authors

Yu-Hsuan Wang, Cheng-Tao Chung, Hung-yi Lee

Topics

Artificial Intelligence > Core AI > Interpretability Deep Learning > Architectures > Neural Networks Speech & Audio > Analysis > Speech Analysis Machine Learning > Learning Types > Deep Learning

Keywords

neural network interpretability speech analysis temporal analysis gated recurrent neural network phoneme boundary gate activation signal phoneme segmentation

More from Interspeech 2017

All Interspeech 2017 papers →

Data Augmentation, Missing Feature Mask and Kernel Classification for Through-the-Wall Acoustic Surveillance

Endpoint Detection Using Grid Long Short-Term Memory Networks for Streaming Speech Recognition

Deep Learning Techniques in Tandem with Signal Processing Cues for Phonetic Segmentation for Text to Speech Synthesis in Indian Languages

Speaker Change Detection in Broadcast TV Using Bidirectional Long Short-Term Memory Networks

Improved Automatic Speech Recognition Using Subband Temporal Envelope Features and Time-Delay Neural Network Denoising Autoencoder

Factored Deep Convolutional Neural Networks for Noise Robust Speech Recognition