Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection (2019-02-13T00:00:00.000000Z)

TL;DR

This model is robust, highly unsupervised, end-to-end and requires minimum preprocessing, feature engineering or hyperparameter tuning and outperforms the state-of-the-art acoustic novelty detectors.

Abstract

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.

Authors

Duong Nguyen

1 papers

O. Kirsebom

1 papers

F. Frazão

1 papers

Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection

TL;DR

Abstract

Authors

References26 items

A Multi-Task Deep Learning Architecture for Maritime Surveillance Using AIS Data Streams

Filtering Variational Objectives

Acoustic novelty detection with adversarial autoencoders

WaveNet: A Generative Model for Raw Audio

Sequential Neural Models with Stochastic Layers

Deep Kalman Filters

A Recurrent Latent Variable Model for Sequential Data

An unsupervised acoustic fall detection system using source separation for sound interference suppression

A novel approach for automatic acoustic novelty detection using a denoising autoencoder with bidirectional LSTM neural networks

Gated Feedback Recurrent Neural Networks

Adam: A Method for Stochastic Optimization

Learning Stochastic Recurrent Networks

The PASCAL CHiME speech separation and recognition challenge

Modeling Temporal Dependencies in High-Dimensional Sequences: Application to Polyphonic Music Generation and Transcription

Probabilistic Novelty Detection for Acoustic Surveillance Under Real-World Conditions

Audio-Visual Fusion for Detecting Violent Scenes in Videos

Acoustic fall detection using Gaussian mixture models and GMM supervectors

CASSANDRA: audio-video sensor fusion for aggression detection

Audio Based Event Detection for Multimedia Surveillance

A Multimodal Audio Visible and Infrared Surveillance System (MAVISS)

Introduction to random signals and applied Kalman filtering, 2nd edn. Robert Grover Brown and Patrick Y. C. Hwang, Wiley, New York, 1992. ISBN 0‐471‐52573‐1, 512 pp., $62.95.

Deep Learning

Long short-term memory recurrent neural network architectures for large scale acoustic modeling

Novelty detection in jet engine vibration spectra

A Tutorial on Particle Filtering and Smoothing: Fifteen years later

Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection van

Field of Study

Journal Information

Name

Page

Venue Information

Name

Type

URL

Alternate Names