Segment-level Metric Learning for Few-shot Bioacoustic Event Detection

TL;DR

This paper introduces a segment-level few-shot learning framework for bioacoustic event detection that leverages both positive and negative examples and uses transductive inference, significantly improving performance over baselines.

Contribution

The study proposes a novel segment-level metric learning approach utilizing negative events and transductive inference to enhance few-shot bioacoustic event detection.

Findings

Achieved an F-measure of 62.73 on DCASE 2022 validation set.

Outperformed the baseline prototypical network by a large margin.

Ranked 2nd in the DCASE2022-T5 challenge.

Abstract

Few-shot bioacoustic event detection is a task that detects the occurrence time of a novel sound given a few examples. Previous methods employ metric learning to build a latent space with the labeled part of different sound classes, also known as positive events. In this study, we propose a segment-level few-shot learning framework that utilizes both the positive and negative events during model optimization. Training with negative events, which are larger in volume than positive events, can increase the generalization ability of the model. In addition, we use transductive inference on the validation set during training for better adaptation to novel classes. We conduct ablation studies on our proposed method with different setups on input features, training data, and hyper-parameters. Our final system achieves an F-measure of 62.73 on the DCASE 2022 challenge task 5 (DCASE2022-T5) validation set, outperforming the performance of the baseline prototypical network 34.02 by a large margin. Using the proposed method, our submitted system ranks 2nd in DCASE2022-T5. The code of this paper is fully open-sourced at https://github.com/haoheliu/DCASE_2022_Task_5.