SoundNet: Learning Sound Representations from Unlabeled Video

TL;DR

SoundNet leverages unlabeled videos and a student-teacher training approach to learn rich sound representations, achieving significant improvements on classification benchmarks and revealing emergent high-level semantics without explicit labels.

Contribution

Introduces a novel method using synchronized video data and student-teacher training to learn sound representations without labeled data.

Findings

Significant performance improvements on acoustic classification benchmarks.

Emergence of high-level semantic features in the sound network.

Effective use of unlabeled video data for sound representation learning.

Abstract

We learn rich natural sound representations by capitalizing on large amounts of unlabeled sound data collected in the wild. We leverage the natural synchronization between vision and sound to learn an acoustic representation using two-million unlabeled videos. Unlabeled video has the advantage that it can be economically acquired at massive scales, yet contains useful signals about natural sound. We propose a student-teacher training procedure which transfers discriminative visual knowledge from well established visual recognition models into the sound modality using unlabeled video as a bridge. Our sound representation yields significant performance improvements over the state-of-the-art results on standard benchmarks for acoustic scene/object classification. Visualizations suggest some high-level semantics automatically emerge in the sound network, even though it is trained without ground truth labels.