A Comparison of Five Multiple Instance Learning Pooling Functions for Sound Event Detection with Weak Labeling

Yun Wang, Juncheng Li, Florian Metze
10/22/2018
cs.SDeess.AS

Official code: MaigoAkisame/cmu-thesis (170 stars, MIT, last commit 2022-05-14).

Abstract

Sound event detection (SED) entails two subtasks: recognizing what types of sound events are present in an audio stream (audio tagging), and pinpointing their onset and offset times (localization). In the popular multiple instance learning (MIL) framework for SED with weak labeling, an important component is the pooling function. This paper compares five types of pooling functions both theoretically and experimentally, with special focus on their performance of localization. Although the attention pooling function is currently receiving the most attention, we find the linear softmax pooling function to perform the best among the five. Using this pooling function, we build a neural network called TALNet. It is the first system to reach state-of-the-art audio tagging performance on Audio Set, while exhibiting strong localization performance on the DCASE 2017 challenge at the same time.

View on arXivView PDF

Code Implementations(1)

MaigoAkisame/cmu-thesisOfficial100%
MIT

Code for Yun Wang's PhD Thesis: Polyphonic Sound Event Detection with Weak Labeling

17047Shell, PythonOct 22, 20184 years agoMIT

Similar Papers

Adaptive pooling operators for weakly labeled sound event detection

88% match1 repo782018

Sound Event Detection and Time-Frequency Segmentation from Weakly Labelled Data

84% match1 repo462018

A Track-Wise Ensemble Event Independent Network for Polyphonic Sound Event Localization and Detection

82% match1 repo242022

Connectionist Temporal Localization for Sound Event Detection with Sequential Labeling

82% match1 repo1702018

An Improved Event-Independent Network for Polyphonic Sound Event Localization and Detection

81% match1 repo812020

Dense-Localizing Audio-Visual Events in Untrimmed Videos: A Large-Scale Benchmark and Baseline

81% match1 repo732023

Cite this paper

@article{wang2018comparison,
  title  = {A Comparison of Five Multiple Instance Learning Pooling Functions for Sound Event Detection with Weak Labeling},
  author = {Yun Wang and Juncheng Li and Florian Metze},
  year   = {2018},
  eprint = {1810.09050},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1810.09050}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1810.09050.svg)](https://codeofpaper.com/paper/1810.09050)

Discussion