Hierarchical learning for DNN-based acoustic scene classification

TL;DR

This paper introduces hierarchical learning methods for DNN-based acoustic scene classification, leveraging taxonomy information to significantly improve performance over traditional GMM benchmarks.

Contribution

It proposes hierarchical pre-training and multi-level objective functions to enhance DNN performance using environmental sound taxonomy.

Findings

22.9% relative error reduction over GMM baseline

Effective incorporation of hierarchical taxonomy improves classification accuracy

Validated on DCASE 2016 challenge dataset

Abstract

In this paper, we present a deep neural network (DNN)-based acoustic scene classification framework. Two hierarchical learning methods are proposed to improve the DNN baseline performance by incorporating the hierarchical taxonomy information of environmental sounds. Firstly, the parameters of the DNN are initialized by the proposed hierarchical pre-training. Multi-level objective function is then adopted to add more constraint on the cross-entropy based loss function. A series of experiments were conducted on the Task1 of the Detection and Classification of Acoustic Scenes and Events (DCASE) 2016 challenge. The final DNN-based system achieved a 22.9% relative improvement on average scene classification error as compared with the Gaussian Mixture Model (GMM)-based benchmark system across four standard folds.