Can I read Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection on EtoBox?

Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection by Nguyen, Duong; Kirsebom, Oliver S.; Frazão, Fábio; Fablet, Ronan; Matwin, Stan is a scholarly article available to read on EtoBox.

What is Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection about?

In this paper, we adapt Recurrent Neural Networks with Stochastic Layers, which are the state-of-the-art for generating text, music and speech, to the problem of acoustic novelty detection. By integrating uncertainty into the hidden states, this type of network is able to learn the distribution of complex sequences. Because the learned distribution can be calculated explicitly in terms of probability, we can evaluate how likely an observation is then detect low-probability events as novel. The model is robust, highly unsupervised, end-to-end and requires minimum preprocessing, feature engineering or hyperparameter tuning. An experiment on a benchmark dataset shows that our model outperforms the state-of-the-art acoustic novelty detectors.

Author

Nguyen, Duong; Kirsebom, Oliver S.; Frazão, Fábio; Fablet, Ronan; Matwin, Stan

Published

2019

Language

EN