Show simple item record

dc.contributor.authorZhang, Dalin
dc.contributor.authorYao, Lina
dc.contributor.authorWang, Sen
dc.contributor.authorChen, Kaixuan
dc.contributor.authorYang, Zheng
dc.contributor.authorBenatallah, Boualem
dc.contributor.editorPhung, D
dc.contributor.editorTseng, VS
dc.contributor.editorWebb, GI
dc.contributor.editorHo, B
dc.contributor.editorGanji, M
dc.contributor.editorRashidi, L
dc.date.accessioned2019-06-09T01:35:29Z
dc.date.available2019-06-09T01:35:29Z
dc.date.issued2018
dc.identifier.isbn978-3-319-93033-6
dc.identifier.issn0302-9743
dc.identifier.doi10.1007/978-3-319-93034-3_13
dc.identifier.urihttp://hdl.handle.net/10072/383657
dc.description.abstractNon-invasive brain-computer interface using electroencephalography (EEG) signals promises a convenient approach empowering humans to communicate with and even control the outside world only with intentions. Herein, we propose to analyze EEG signals using fuzzy integral with deep reinforcement learning optimization to aggregate two aspects of information contained within EEG signals, namely local spatio-temporal and global temporal information, and demonstrate its benefits in EEG-based human intention recognition tasks. The EEG signals are first transformed into a 3D format preserving both topological and temporal structures, followed by distinctive local spatio-temporal feature extraction by a 3D-CNN, as well as the global temporal feature extraction by an RNN. Next, a fuzzy integral with respect to the optimized fuzzy measures with deep reinforcement learning is utilized to integrate the two extracted information and makes a final decision. The proposed approach retains the topological and temporal structures of EEG signals and merges them in a more efficient way. Experiments on a public EEG-based movement intention dataset demonstrate the effectiveness and superior performance of our proposed method.
dc.description.peerreviewedYes
dc.languageEnglish
dc.publisherSPRINGER INTERNATIONAL PUBLISHING AG
dc.relation.ispartofconferencename22nd Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD)
dc.relation.ispartofconferencetitleADVANCES IN KNOWLEDGE DISCOVERY AND DATA MINING, PAKDD 2018, PT I
dc.relation.ispartofdatefrom2018-06-03
dc.relation.ispartofdateto2018-06-06
dc.relation.ispartoflocationDeakin Univ, Melbourne, AUSTRALIA
dc.relation.ispartofpagefrom156
dc.relation.ispartofpageto168
dc.relation.ispartofvolume10937
dc.subject.fieldofresearchArtificial intelligence
dc.subject.fieldofresearchcode4602
dc.titleFuzzy Integral Optimization with Deep Q-Network for EEG-Based Intention Recognition
dc.typeConference output
dc.type.descriptionE1 - Conferences
dc.type.codeE - Conference Publications
gro.hasfulltextNo Full Text
gro.griffith.authorWang, Sen


Files in this item

FilesSizeFormatView

There are no files associated with this item.

This item appears in the following Collection(s)

  • Conference outputs
    Contains papers delivered by Griffith authors at national and international conferences.

Show simple item record