Play It Back: Iterative Attention for Audio Recognition

TL;DR

This paper introduces an attention-based model that iteratively replays and refines discriminative audio segments to improve classification accuracy, achieving state-of-the-art results on multiple benchmarks.

Contribution

The proposed model employs iterative attention and selective replay of audio segments, a novel approach enhancing fine-grained audio recognition performance.

Findings

Achieves state-of-the-art accuracy on AudioSet, VGG-Sound, and EPIC-KITCHENS-100.

Effectively refines focus on discriminative sounds through iterative replay.

Demonstrates the benefit of selective segment replay for audio classification.

Abstract

A key function of auditory cognition is the association of characteristic sounds with their corresponding semantics over time. Humans attempting to discriminate between fine-grained audio categories, often replay the same discriminative sounds to increase their prediction confidence. We propose an end-to-end attention-based architecture that through selective repetition attends over the most discriminative sounds across the audio sequence. Our model initially uses the full audio sequence and iteratively refines the temporal segments replayed based on slot attention. At each playback, the selected segments are replayed using a smaller hop length which represents higher resolution features within these segments. We show that our method can consistently achieve state-of-the-art performance across three audio-classification benchmarks: AudioSet, VGG-Sound, and EPIC-KITCHENS-100.