Weakly Labeled Learning Using BLSTM-CTC for Sound Event Detection

Taiki Matsuyoshi, Tatsuya Komatsu, Reishi Kondo, Takeshi Yamada, Shoji Makino

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

In this paper, we propose a method of weakly labeled learning of bidirectional long short-term memory (BLSTM) using connectionist temporal classification (BLSTM-CTC) to reduce the hand-labeling cost of learning samples. BLSTM-CTC enables us to update the parameters of BLSTM by loss calculation using CTC, instead of the exact error calculation that cannot be conducted when using weakly labeled samples, which have only the event class of each individual sound event. In the proposed method, we first conduct strongly labeled learning of BLSTM using a small amount of strongly labeled samples, which have the timestamps of the beginning and end of each individual sound event and its event class, as initial learning. We then conduct weakly labeled learning based on BLSTM-CTC using a large amount of weakly labeled samples as additional learning. To evaluate the performance of the proposed method, we conducted a sound event detection experiment using the dataset provided by Detection and Classification of Acoustic Scenes and Events (DCASE) 2016 Task 2. As a result, the proposed method improved the segment-based F1 score by 1.9% compared with the initial learning mentioned above. Furthermore, it succeeded in reducing the labeling cost by 95%, although the F1 score was degraded by 1.3%, comparing with additional learning using a large amount of strongly labeled samples. This result confirms that our weakly labeled learning is effective for learning BLSTM with a low hand-labeling cost.

Original languageEnglish
Title of host publication2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1918-1923
Number of pages6
ISBN (Electronic)9789881476852
DOIs
Publication statusPublished - 2019 Mar 4
Externally publishedYes
Event10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Honolulu, United States
Duration: 2018 Nov 122018 Nov 15

Publication series

Name2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings

Conference

Conference10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018
CountryUnited States
CityHonolulu
Period18/11/1218/11/15

ASJC Scopus subject areas

  • Information Systems

Fingerprint Dive into the research topics of 'Weakly Labeled Learning Using BLSTM-CTC for Sound Event Detection'. Together they form a unique fingerprint.

Cite this