Enriching Ontology with Temporal Commonsense for Low-Resource Audio Tagging

TL;DR

This paper enhances low-resource audio tagging by integrating temporal commonsense knowledge through a semi-automatic construction of temporal knowledge graphs and a relation-aware graph neural network, improving performance on urban sound datasets.

Contribution

It introduces a novel semi-automatic method to construct temporal knowledge graphs and a D-GCN model to effectively combine ontological and temporal knowledge for audio tagging.

Findings

Temporal knowledge improves tagging accuracy.

Combined knowledge graphs outperform single-source models.

Effective in low-resource scenarios.

Abstract

Audio tagging aims at predicting sound events occurred in a recording. Traditional models require enormous laborious annotations, otherwise performance degeneration will be the norm. Therefore, we investigate robust audio tagging models in low-resource scenarios with the enhancement of knowledge graphs. Besides existing ontological knowledge, we further propose a semi-automatic approach that can construct temporal knowledge graphs on diverse domain-specific label sets. Moreover, we leverage a variant of relation-aware graph neural network, D-GCN, to combine the strength of the two knowledge types. Experiments on AudioSet and SONYC urban sound tagging datasets suggest the effectiveness of the introduced temporal knowledge, and the advantage of the combined KGs with D-GCN over single knowledge source.