General-purpose audio tagging using deep learning approaches

Abstract

In this talk, I will explain the techniques and models applied to the submission of Cochlear.ai team for DCASE 2018 task 2: General-purpose audio tagging of Freesound content with AudioSet labels. We mainly focused on how to train deep learning models efficiently against strong augmentation and label noise. First, we conducted a single-block DenseNet architecture and multi-head softmax classifier for efficient learning with mixup augmentation. For the label noise, we applied the batch-wise loss masking to eliminate the loss of outliers in a mini-batch. We also tried an ensemble of various models, trained by using different sampling rate or audio representation.

Biography

Il-Young Jeong is a co-founder and research scientist at Cochlear.ai. He holds a B.S. degree in Electrical Engineering and received M.E. from Intelligent Information Processing laboratory, Sogang University. His primary interest was digital signal processing for speech recognition systems. During his PhD study at the Music and Audio Research Group, Seoul National University, he sought to expand his research area into general audio signals such as music and sound events.