WEKO3
アイテム
{"_buckets": {"deposit": "556b138c-cea5-49b5-ad6a-01d38c9c9295"}, "_deposit": {"created_by": 30, "id": "1668", "owners": [30], "pid": {"revision_id": 0, "type": "depid", "value": "1668"}, "status": "published"}, "_oai": {"id": "oai:oist.repo.nii.ac.jp:00001668", "sets": ["78"]}, "author_link": ["10012", "10011"], "item_10001_biblio_info_7": {"attribute_name": "Bibliographic Information", "attribute_value_mlt": [{"bibliographicIssueDates": {"bibliographicIssueDate": "2020-05-18", "bibliographicIssueDateType": "Issued"}, "bibliographicIssueNumber": "5", "bibliographicPageStart": "564", "bibliographicVolumeNumber": "22", "bibliographic_titles": [{}, {"bibliographic_title": "Entropy", "bibliographic_titleLang": "en"}]}]}, "item_10001_creator_3": {"attribute_name": "Author", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "Matsumoto, Takazumi"}], "nameIdentifiers": [{"nameIdentifier": "10011", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "Tani, Jun"}], "nameIdentifiers": [{"nameIdentifier": "10012", "nameIdentifierScheme": "WEKO"}]}]}, "item_10001_description_5": {"attribute_name": "Abstract", "attribute_value_mlt": [{"subitem_description": "It is crucial to ask how agents can achieve goals by generating action plans using only partial models of the world acquired through habituated sensory-motor experiences. Although many existing robotics studies use a forward model framework, there are generalization issues with high degrees of freedom. The current study shows that the predictive coding (PC) and active inference (AIF) frameworks, which employ a generative model, can develop better generalization by learning a prior distribution in a low dimensional latent state space representing probabilistic structures extracted from well habituated sensory-motor trajectories. In our proposed model, learning is carried out by inferring optimal latent variables as well as synaptic weights for maximizing the evidence lower bound, while goal-directed planning is accomplished by inferring latent variables for maximizing the estimated lower bound. Our proposed model was evaluated with both simple and complex robotic tasks in simulation, which demonstrated sufficient generalization in learning with limited training data by setting an intermediate value for a regularization coefficient. Furthermore, comparative simulation results show that the proposed model outperforms a conventional forward model in goal-directed planning, due to the learned prior confining the search of motor plans within the range of habituated trajectories. ", "subitem_description_type": "Other"}]}, "item_10001_publisher_8": {"attribute_name": "Publisher", "attribute_value_mlt": [{"subitem_publisher": "MDPI "}]}, "item_10001_relation_14": {"attribute_name": "DOI", "attribute_value_mlt": [{"subitem_relation_type": "isIdenticalTo", "subitem_relation_type_id": {"subitem_relation_type_id_text": "info:doi/10.3390/e22050564", "subitem_relation_type_select": "DOI"}}]}, "item_10001_relation_17": {"attribute_name": "Related site", "attribute_value_mlt": [{"subitem_relation_type_id": {"subitem_relation_type_id_text": "https://www.mdpi.com/1099-4300/22/5/564/htm", "subitem_relation_type_select": "URI"}}]}, "item_10001_rights_15": {"attribute_name": "Rights", "attribute_value_mlt": [{"subitem_rights": "© 2020 The Author(s)"}]}, "item_10001_source_id_9": {"attribute_name": "ISSN", "attribute_value_mlt": [{"subitem_source_identifier": "1099-4300", "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_970fb48d4fbd8a85", "subitem_version_type": "VoR"}]}, "item_files": {"attribute_name": "ファイル情報", "attribute_type": "file", "attribute_value_mlt": [{"accessrole": "open_date", "date": [{"dateType": "Available", "dateValue": "2020-08-06"}], "displaytype": "detail", "download_preview_message": "", "file_order": 0, "filename": "Matsumoto-2020-Goal-Directed Planning for Habi.pdf", "filesize": [{"value": "8.0 MB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensefree": "Creative Commons Attribution 4.0 International(https://creativecommons.org/licenses/by/4.0/)", "licensetype": "license_free", "mimetype": "application/pdf", "size": 8000000.0, "url": {"label": "Matsumoto-2020-Goal-Directed Planning for Habi", "url": "https://oist.repo.nii.ac.jp/record/1668/files/Matsumoto-2020-Goal-Directed Planning for Habi.pdf"}, "version_id": "dc4a0b3e-72eb-4912-948d-0d658fb24868"}]}, "item_keyword": {"attribute_name": "キーワード", "attribute_value_mlt": [{"subitem_subject": "goal directed planning", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "active inference", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "predictive coding", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "variational bayes", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "recurrent neural network", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}]}, "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": "Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network", "item_titles": {"attribute_name": "タイトル", "attribute_value_mlt": [{"subitem_title": "Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network", "subitem_title_language": "en"}]}, "item_type_id": "10001", "owner": "30", "path": ["78"], "permalink_uri": "https://oist.repo.nii.ac.jp/records/1668", "pubdate": {"attribute_name": "公開日", "attribute_value": "2020-08-06"}, "publish_date": "2020-08-06", "publish_status": "0", "recid": "1668", "relation": {}, "relation_version_is_last": true, "title": ["Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network"], "weko_shared_id": 30}
Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network
https://oist.repo.nii.ac.jp/records/1668
https://oist.repo.nii.ac.jp/records/1668cd4caa8a-c92a-44fc-b851-bc14bdb217bf
名前 / ファイル | ライセンス | アクション |
---|---|---|
Matsumoto-2020-Goal-Directed Planning for Habi (8.0 MB)
|
Creative Commons Attribution 4.0 International(https://creativecommons.org/licenses/by/4.0/)
|
Item type | 学術雑誌論文 / Journal Article(1) | |||||
---|---|---|---|---|---|---|
公開日 | 2020-08-06 | |||||
タイトル | ||||||
言語 | en | |||||
タイトル | Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network | |||||
言語 | ||||||
言語 | eng | |||||
キーワード | ||||||
言語 | en | |||||
主題Scheme | Other | |||||
主題 | goal directed planning | |||||
キーワード | ||||||
言語 | en | |||||
主題Scheme | Other | |||||
主題 | active inference | |||||
キーワード | ||||||
言語 | en | |||||
主題Scheme | Other | |||||
主題 | predictive coding | |||||
キーワード | ||||||
言語 | en | |||||
主題Scheme | Other | |||||
主題 | variational bayes | |||||
キーワード | ||||||
言語 | en | |||||
主題Scheme | Other | |||||
主題 | recurrent neural network | |||||
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) |
Matsumoto, Takazumi
× Matsumoto, Takazumi× Tani, Jun |
|||||
書誌情報 |
en : Entropy 巻 22, 号 5, p. 564, 発行日 2020-05-18 |
|||||
抄録 | ||||||
内容記述タイプ | Other | |||||
内容記述 | It is crucial to ask how agents can achieve goals by generating action plans using only partial models of the world acquired through habituated sensory-motor experiences. Although many existing robotics studies use a forward model framework, there are generalization issues with high degrees of freedom. The current study shows that the predictive coding (PC) and active inference (AIF) frameworks, which employ a generative model, can develop better generalization by learning a prior distribution in a low dimensional latent state space representing probabilistic structures extracted from well habituated sensory-motor trajectories. In our proposed model, learning is carried out by inferring optimal latent variables as well as synaptic weights for maximizing the evidence lower bound, while goal-directed planning is accomplished by inferring latent variables for maximizing the estimated lower bound. Our proposed model was evaluated with both simple and complex robotic tasks in simulation, which demonstrated sufficient generalization in learning with limited training data by setting an intermediate value for a regularization coefficient. Furthermore, comparative simulation results show that the proposed model outperforms a conventional forward model in goal-directed planning, due to the learned prior confining the search of motor plans within the range of habituated trajectories. | |||||
出版者 | ||||||
出版者 | MDPI | |||||
ISSN | ||||||
収録物識別子タイプ | ISSN | |||||
収録物識別子 | 1099-4300 | |||||
DOI | ||||||
関連タイプ | isIdenticalTo | |||||
識別子タイプ | DOI | |||||
関連識別子 | info:doi/10.3390/e22050564 | |||||
権利 | ||||||
権利情報 | © 2020 The Author(s) | |||||
関連サイト | ||||||
識別子タイプ | URI | |||||
関連識別子 | https://www.mdpi.com/1099-4300/22/5/564/htm | |||||
著者版フラグ | ||||||
出版タイプ | VoR | |||||
出版タイプResource | http://purl.org/coar/version/c_970fb48d4fbd8a85 |