Surrey-cvssp system for DCASE2017 challenge task4

TL;DR

This paper presents a system for the DCASE2017 challenge task 4, utilizing CNN and GRU neural networks with novel gating and attention mechanisms, achieving significant improvements over baseline methods in sound event detection and audio tagging.

Contribution

The paper introduces a learnable gating activation function, an attention-based localization scheme, and a batch-level balancing strategy for weakly labeled sound event detection.

Findings

Achieved 61% F-value in audio tagging

Achieved 0.73 error rate in sound event detection

Outperformed baseline MLP system significantly

Abstract

In this technique report, we present a bunch of methods for the task 4 of Detection and Classification of Acoustic Scenes and Events 2017 (DCASE2017) challenge. This task evaluates systems for the large-scale detection of sound events using weakly labeled training data. The data are YouTube video excerpts focusing on transportation and warnings due to their industry applications. There are two tasks, audio tagging and sound event detection from weakly labeled data. Convolutional neural network (CNN) and gated recurrent unit (GRU) based recurrent neural network (RNN) are adopted as our basic framework. We proposed a learnable gating activation function for selecting informative local features. Attention-based scheme is used for localizing the specific events in a weakly-supervised mode. A new batch-level balancing strategy is also proposed to tackle the data unbalancing problem. Fusion of posteriors from different systems are found effective to improve the performance. In a summary, we get 61% F-value for the audio tagging subtask and 0.73 error rate (ER) for the sound event detection subtask on the development set. While the official multilayer perceptron (MLP) based baseline just obtained 13.1% F-value for the audio tagging and 1.02 for the sound event detection.