Dynamic selective auditory attention detection using RNN and reinforcement learning

Abstract The cocktail party phenomenon describes the ability of the human brain to focus auditory attention on a particular stimulus while ignoring other acoustic events. Selective auditory attention detection (SAAD) is an important issue in the development of brain-computer interface systems and co...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Masoud Geravanchizadeh, Hossein Roushan
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2021
Materias:
R
Q
Acceso en línea:https://doaj.org/article/61a6925d5c0242f5a2e7ad00979805c7
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:61a6925d5c0242f5a2e7ad00979805c7
record_format dspace
spelling oai:doaj.org-article:61a6925d5c0242f5a2e7ad00979805c72021-12-02T16:31:52ZDynamic selective auditory attention detection using RNN and reinforcement learning10.1038/s41598-021-94876-02045-2322https://doaj.org/article/61a6925d5c0242f5a2e7ad00979805c72021-07-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-94876-0https://doaj.org/toc/2045-2322Abstract The cocktail party phenomenon describes the ability of the human brain to focus auditory attention on a particular stimulus while ignoring other acoustic events. Selective auditory attention detection (SAAD) is an important issue in the development of brain-computer interface systems and cocktail party processors. This paper proposes a new dynamic attention detection system to process the temporal evolution of the input signal. The proposed dynamic SAAD is modeled as a sequential decision-making problem, which is solved by recurrent neural network (RNN) and reinforcement learning methods of Q-learning and deep Q-learning. Among different dynamic learning approaches, the evaluation results show that the deep Q-learning approach with RNN as agent provides the highest classification accuracy (94.2%) with the least detection delay. The proposed SAAD system is advantageous, in the sense that the detection of attention is performed dynamically for the sequential inputs. Also, the system has the potential to be used in scenarios, where the attention of the listener might be switched in time in the presence of various acoustic events.Masoud GeravanchizadehHossein RoushanNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-11 (2021)
institution DOAJ
collection DOAJ
language EN
topic Medicine
R
Science
Q
spellingShingle Medicine
R
Science
Q
Masoud Geravanchizadeh
Hossein Roushan
Dynamic selective auditory attention detection using RNN and reinforcement learning
description Abstract The cocktail party phenomenon describes the ability of the human brain to focus auditory attention on a particular stimulus while ignoring other acoustic events. Selective auditory attention detection (SAAD) is an important issue in the development of brain-computer interface systems and cocktail party processors. This paper proposes a new dynamic attention detection system to process the temporal evolution of the input signal. The proposed dynamic SAAD is modeled as a sequential decision-making problem, which is solved by recurrent neural network (RNN) and reinforcement learning methods of Q-learning and deep Q-learning. Among different dynamic learning approaches, the evaluation results show that the deep Q-learning approach with RNN as agent provides the highest classification accuracy (94.2%) with the least detection delay. The proposed SAAD system is advantageous, in the sense that the detection of attention is performed dynamically for the sequential inputs. Also, the system has the potential to be used in scenarios, where the attention of the listener might be switched in time in the presence of various acoustic events.
format article
author Masoud Geravanchizadeh
Hossein Roushan
author_facet Masoud Geravanchizadeh
Hossein Roushan
author_sort Masoud Geravanchizadeh
title Dynamic selective auditory attention detection using RNN and reinforcement learning
title_short Dynamic selective auditory attention detection using RNN and reinforcement learning
title_full Dynamic selective auditory attention detection using RNN and reinforcement learning
title_fullStr Dynamic selective auditory attention detection using RNN and reinforcement learning
title_full_unstemmed Dynamic selective auditory attention detection using RNN and reinforcement learning
title_sort dynamic selective auditory attention detection using rnn and reinforcement learning
publisher Nature Portfolio
publishDate 2021
url https://doaj.org/article/61a6925d5c0242f5a2e7ad00979805c7
work_keys_str_mv AT masoudgeravanchizadeh dynamicselectiveauditoryattentiondetectionusingrnnandreinforcementlearning
AT hosseinroushan dynamicselectiveauditoryattentiondetectionusingrnnandreinforcementlearning
_version_ 1718383811879763968