Few-shot bioacoustic event detection at the DCASE 2023 challenge

TL;DR

This paper discusses the 2023 DCASE challenge on few-shot bioacoustic event detection, highlighting task evolution, increased complexity, and improved F-scores without ensemble models.

Contribution

It provides an overview of the challenge's evolution, new rules, and the performance of recent systems, emphasizing advancements in few-shot bioacoustic detection.

Findings

F-scores reached up to 63% on evaluation set

Participation increased with 6 teams submitting

Systems have become more complex over time

Abstract

Few-shot bioacoustic event detection consists in detecting sound events of specified types, in varying soundscapes, while having access to only a few examples of the class of interest. This task ran as part of the DCASE challenge for the third time this year with an evaluation set expanded to include new animal species, and a new rule: ensemble models were no longer allowed. The 2023 few shot task received submissions from 6 different teams with F-scores reaching as high as 63% on the evaluation set. Here we describe the task, focusing on describing the elements that differed from previous years. We also take a look back at past editions to describe how the task has evolved. Not only have the F-score results steadily improved (40% to 60% to 63%), but the type of systems proposed have also become more complex. Sound event detection systems are no longer simple variations of the baselines provided: multiple few-shot learning methodologies are still strong contenders for the task.