1 |
Weakly Labelled AudioSet Tagging With Attention Neural Networks
|
|
|
|
Abstract:
Audio tagging is the task of predicting the presence or absence of sound classes within an audio clip. Previous work in audio tagging focused on relatively small datasets limited to recognising a small number of sound classes. We investigate audio tagging on AudioSet, which is a dataset consisting of over 2 million audio clips and 527 classes. AudioSet is weakly labelled, in that only the presence or absence of sound classes is known for each clip, while the onset and offset times are unknown. To address the weakly-labelled audio tagging problem, we propose attention neural networks as a way to attend the most salient parts of an audio clip. We bridge the connection between attention neural networks and multiple instance learning (MIL) methods, and propose decision-level and feature-level attention neural networks for audio tagging. We investigate attention neural networks modelled by different functions, depths and widths. Experiments on AudioSet show that the feature-level attention neural network achieves a state-of-the-art mean average precision (mAP) of 0.369, outperforming the best multiple instance learning (MIL) method of 0.317 and Google’s deep neural network baseline of 0.314. In addition, we discover that the audio tagging performance on AudioSet embedding features has a weak correlation with the number of training examples and the quality of labels of each sound class.
|
|
URL: https://doi.org/10.1109/TASLP.2019.2930913 http://epubs.surrey.ac.uk/852511/1/KongYuXuIWP19-aslp-audioset_accepted_ieee.pdf
|
|
BASE
|
|
Hide details
|
|
2 |
LaSOT: A High-quality Benchmark for Large-scale Single Object Tracking ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
Novel elicitation and annotation schemes for sentential and sub-sentential alignments of bitexts
|
|
|
|
In: Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC'16) ; Language Resources and Evaluation Conference (LREC'16) ; https://hal.archives-ouvertes.fr/hal-03396226 ; Language Resources and Evaluation Conference (LREC'16), ELRA, May 2016, Portoroz, Slovenia ; http://lrec2016.lrec-conf.org/en/ (2016)
|
|
BASE
|
|
Show details
|
|
4 |
A 2D CRF Model for Sentence Alignment
|
|
|
|
In: 9th Workshop on Building and Using Comparable Corpora ; https://hal.archives-ouvertes.fr/hal-01388656 ; 9th Workshop on Building and Using Comparable Corpora, 2016, Portorož, Slovenia ; http://lrec2016.lrec-conf.org/en/ (2016)
|
|
BASE
|
|
Show details
|
|
5 |
Confidence Measures for Alignment and for Machine Translation ; Mesures de Confiance pour l’Alignement et pour la Traduction Automatique
|
|
|
|
In: https://tel.archives-ouvertes.fr/tel-01399222 ; Signal and Image Processing. Université Paris Saclay (COmUE), 2016. English. ⟨NNT : 2016SACLS270⟩ (2016)
|
|
BASE
|
|
Show details
|
|
6 |
TransRead: Designing a Bilingual Reading Experience with Machine Translation Technologies
|
|
|
|
In: NAACL ; https://hal.inria.fr/hal-01370497 ; NAACL, North American Chapter of the Association for Computational Linguistics, Jun 2016, San Diego, United States. pp.5, ⟨10.18653/v1/N16-3006⟩ ; naacl.org/naacl-hlt-2016/ (2016)
|
|
BASE
|
|
Show details
|
|
7 |
Deep Neural Network for Robust Speech Recognition With Auxiliary Features From Laser-Doppler Vibrometer Sensor
|
|
|
|
BASE
|
|
Show details
|
|
8 |
Sentence Alignment for Literary Texts
|
|
|
|
In: ISSN: 1945-3604 ; Linguistic Issues in Language Technology ; https://hal.archives-ouvertes.fr/hal-01634995 ; Linguistic Issues in Language Technology, Stanford Calif.: CSLI Publications, 2015, 12, pp.1-25 (2015)
|
|
BASE
|
|
Show details
|
|
10 |
Evaluation of a new automated spotter style exam for assessment of anatomical knowledge
|
|
|
|
BASE
|
|
Show details
|
|
|
|