WEKO3
アイテム
{"_buckets": {"deposit": "7ef823ae-aeaf-459a-b314-7c243c4c1421"}, "_deposit": {"created_by": 27, "id": "951", "owners": [27], "pid": {"revision_id": 0, "type": "depid", "value": "951"}, "status": "published"}, "_oai": {"id": "oai:oist.repo.nii.ac.jp:00000951", "sets": ["78"]}, "author_link": ["5225", "5226", "5224"], "item_10001_biblio_info_7": {"attribute_name": "Bibliographic Information", "attribute_value_mlt": [{"bibliographicIssueDates": {"bibliographicIssueDate": "2018-05-22", "bibliographicIssueDateType": "Issued"}, "bibliographicPageEnd": "370", "bibliographicPageStart": "356", "bibliographicVolumeNumber": "105", "bibliographic_titles": [{}, {"bibliographic_title": "Neural Networks", "bibliographic_titleLang": "en"}]}]}, "item_10001_creator_3": {"attribute_name": "Author", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "Jung, Minju"}], "nameIdentifiers": [{"nameIdentifier": "5224", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Lee, Haanvid"}], "nameIdentifiers": [{"nameIdentifier": "5225", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Tani, Jun"}], "nameIdentifiers": [{"nameIdentifier": "5226", "nameIdentifierScheme": "WEKO"}]}]}, "item_10001_description_5": {"attribute_name": "Abstract", "attribute_value_mlt": [{"subitem_description": "Video image recognition has been extensively studied with rapid progress recently. However, most methods focus on short-term rather than long-term (contextual) video recognition. Convolutional recurrent neural networks (ConvRNNs) provide robust spatio-temporal information processing capabilities for contextual video recognition, but require extensive computation that slows down training. Inspired by normalization and detrending methods, in this paper we propose \"adaptive detrending\" (AD) for temporal normalization in order to accelerate the training of ConvRNNs, especially of convolutional gated recurrent unit (ConvGRU). For each neuron in a recurrent neural network (RNN), AD identifies the trending change within a sequence and subtracts it, removing the internal covariate shift. In experiments testing for contextual video recognition with ConvGRU, results show that (1) ConvGRU clearly outperforms feed-forward neural networks, (2) AD consistently and significantly accelerates training and improves generalization, (3) performance is further improved when AD is coupled with other normalization methods, and most importantly, (4) the more long-term contextual information is required, the more AD outperforms existing methods.", "subitem_description_type": "Other"}]}, "item_10001_publisher_8": {"attribute_name": "Publisher", "attribute_value_mlt": [{"subitem_publisher": "Elsevier Ltd."}]}, "item_10001_relation_13": {"attribute_name": "PubMedNo.", "attribute_value_mlt": [{"subitem_relation_type": "isVersionOf", "subitem_relation_type_id": {"subitem_relation_type_id_text": "info:pmid/29936360", "subitem_relation_type_select": "PMID"}}]}, "item_10001_relation_14": {"attribute_name": "DOI", "attribute_value_mlt": [{"subitem_relation_type": "isVersionOf", "subitem_relation_type_id": {"subitem_relation_type_id_text": "info:doi/10.1016/j.neunet.2018.05.009", "subitem_relation_type_select": "DOI"}}]}, "item_10001_relation_16": {"attribute_name": "情報源", "attribute_value_mlt": [{"subitem_relation_name": [{"subitem_relation_name_text": "https://creativecommons.org/licenses/by-nc-nd/4.0/"}]}]}, "item_10001_relation_17": {"attribute_name": "Related site", "attribute_value_mlt": [{"subitem_relation_type_id": {"subitem_relation_type_id_text": "https://www.sciencedirect.com/science/article/pii/S0893608018301710?via%3Dihub", "subitem_relation_type_select": "URI"}}]}, "item_10001_rights_15": {"attribute_name": "Rights", "attribute_value_mlt": [{"subitem_rights": "© 2018 Elsevier Ltd."}]}, "item_10001_source_id_9": {"attribute_name": "ISSN", "attribute_value_mlt": [{"subitem_source_identifier": "0893-6080", "subitem_source_identifier_type": "ISSN"}, {"subitem_source_identifier": "1879-2782", "subitem_source_identifier_type": "ISSN"}]}, "item_10001_version_type_20": {"attribute_name": "Author\u0027s flag", "attribute_value_mlt": [{"subitem_version_resource": "http://purl.org/coar/version/c_ab4af688f83e57aa", "subitem_version_type": "AM"}]}, "item_files": {"attribute_name": "ファイル情報", "attribute_type": "file", "attribute_value_mlt": [{"accessrole": "open_date", "date": [{"dateType": "Available", "dateValue": "2020-05-23"}], "displaytype": "detail", "download_preview_message": "", "file_order": 0, "filename": "Accepted Manuscript CC-BY-NC-ND Jung-2018.pdf", "filesize": [{"value": "1.4 MB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensefree": "Creative Commons\nAttribution-NonCommercial-NoDerivatives 4.0 International\n(https://creativecommons.org/licenses/by-nc-nd/4.0/)", "licensetype": "license_free", "mimetype": "application/pdf", "size": 1400000.0, "url": {"label": "Accepted Manuscript CC-BY-NC-ND Jung-2018", "url": "https://oist.repo.nii.ac.jp/record/951/files/Accepted Manuscript CC-BY-NC-ND Jung-2018.pdf"}, "version_id": "0184e0a5-bf8a-477a-831c-3dd2171f6bd0"}]}, "item_language": {"attribute_name": "言語", "attribute_value_mlt": [{"subitem_language": "eng"}]}, "item_resource_type": {"attribute_name": "資源タイプ", "attribute_value_mlt": [{"resourcetype": "journal article", "resourceuri": "http://purl.org/coar/resource_type/c_6501"}]}, "item_title": "Adaptive detrending to accelerate convolutional gated recurrent unit training for contextual video recognition", "item_titles": {"attribute_name": "タイトル", "attribute_value_mlt": [{"subitem_title": "Adaptive detrending to accelerate convolutional gated recurrent unit training for contextual video recognition", "subitem_title_language": "en"}]}, "item_type_id": "10001", "owner": "27", "path": ["78"], "permalink_uri": "https://oist.repo.nii.ac.jp/records/951", "pubdate": {"attribute_name": "公開日", "attribute_value": "2019-06-18"}, "publish_date": "2019-06-18", "publish_status": "0", "recid": "951", "relation": {}, "relation_version_is_last": true, "title": ["Adaptive detrending to accelerate convolutional gated recurrent unit training for contextual video recognition"], "weko_shared_id": 27}
Adaptive detrending to accelerate convolutional gated recurrent unit training for contextual video recognition
https://oist.repo.nii.ac.jp/records/951
https://oist.repo.nii.ac.jp/records/951e05054aa-d252-4c0f-92ac-48731c8d5f10
名前 / ファイル | ライセンス | アクション |
---|---|---|
Accepted Manuscript CC-BY-NC-ND Jung-2018 (1.4 MB)
|
Creative Commons
Attribution-NonCommercial-NoDerivatives 4.0 International (https://creativecommons.org/licenses/by-nc-nd/4.0/) |
Item type | 学術雑誌論文 / Journal Article(1) | |||||
---|---|---|---|---|---|---|
公開日 | 2019-06-18 | |||||
タイトル | ||||||
言語 | en | |||||
タイトル | Adaptive detrending to accelerate convolutional gated recurrent unit training for contextual video recognition | |||||
言語 | ||||||
言語 | eng | |||||
資源タイプ | ||||||
資源タイプ識別子 | http://purl.org/coar/resource_type/c_6501 | |||||
資源タイプ | journal article | |||||
著者(英) |
Jung, Minju
× Jung, Minju× Lee, Haanvid× Tani, Jun |
|||||
書誌情報 |
en : Neural Networks 巻 105, p. 356-370, 発行日 2018-05-22 |
|||||
抄録 | ||||||
内容記述タイプ | Other | |||||
内容記述 | Video image recognition has been extensively studied with rapid progress recently. However, most methods focus on short-term rather than long-term (contextual) video recognition. Convolutional recurrent neural networks (ConvRNNs) provide robust spatio-temporal information processing capabilities for contextual video recognition, but require extensive computation that slows down training. Inspired by normalization and detrending methods, in this paper we propose "adaptive detrending" (AD) for temporal normalization in order to accelerate the training of ConvRNNs, especially of convolutional gated recurrent unit (ConvGRU). For each neuron in a recurrent neural network (RNN), AD identifies the trending change within a sequence and subtracts it, removing the internal covariate shift. In experiments testing for contextual video recognition with ConvGRU, results show that (1) ConvGRU clearly outperforms feed-forward neural networks, (2) AD consistently and significantly accelerates training and improves generalization, (3) performance is further improved when AD is coupled with other normalization methods, and most importantly, (4) the more long-term contextual information is required, the more AD outperforms existing methods. | |||||
出版者 | ||||||
出版者 | Elsevier Ltd. | |||||
ISSN | ||||||
収録物識別子タイプ | ISSN | |||||
収録物識別子 | 0893-6080 | |||||
ISSN | ||||||
収録物識別子タイプ | ISSN | |||||
収録物識別子 | 1879-2782 | |||||
PubMed番号 | ||||||
関連タイプ | isVersionOf | |||||
識別子タイプ | PMID | |||||
関連識別子 | info:pmid/29936360 | |||||
DOI | ||||||
関連タイプ | isVersionOf | |||||
識別子タイプ | DOI | |||||
関連識別子 | info:doi/10.1016/j.neunet.2018.05.009 | |||||
権利 | ||||||
権利情報 | © 2018 Elsevier Ltd. | |||||
情報源 | ||||||
関連名称 | https://creativecommons.org/licenses/by-nc-nd/4.0/ | |||||
関連サイト | ||||||
識別子タイプ | URI | |||||
関連識別子 | https://www.sciencedirect.com/science/article/pii/S0893608018301710?via%3Dihub | |||||
著者版フラグ | ||||||
出版タイプ | AM | |||||
出版タイプResource | http://purl.org/coar/version/c_ab4af688f83e57aa |