{"created":"2025-01-19T01:17:05.076298+00:00","updated":"2025-01-19T15:49:56.952684+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00216508","sets":["1164:3616:10863:10864"]},"path":["10864"],"owner":"44499","recid":"216508","title":["Moving Closer to Real-world Reinforcement Learning"],"pubdate":{"attribute_name":"公開日","attribute_value":"2022-02-18"},"_buckets":{"deposit":"3c207cd5-df3a-41d7-8f58-2ae9d8583f08"},"_deposit":{"id":"216508","pid":{"type":"depid","value":"216508","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"Moving Closer to Real-world Reinforcement Learning","author_link":["558911","558910"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Moving Closer to Real-world Reinforcement Learning"},{"subitem_title":"Moving Closer to Real-world Reinforcement Learning","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"招待講演","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2022-02-18","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"Google Brain"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Google Brain","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/216508/files/IPSJ-AVM22116009.pdf","label":"IPSJ-AVM22116009.pdf"},"date":[{"dateType":"Available","dateValue":"2024-02-18"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-AVM22116009.pdf","filesize":[{"value":"562.6 kB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"5"},{"tax":["include_tax"],"price":"0","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"27"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"7655cbb5-9e44-49d3-bae7-340a8982046c","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2022 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Yunjin, Tang"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Yunjin, Tang","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10438399","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8582","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"From super-human performance in games to ultra-efficient automatic chip layout design, we have witnessed the rise of deep reinforcement learning (RL) and its transition from proof-of-concept tasks to beneficial real world applications. In this presentation, I will show some research topics that are critical for us to further advance toward more and more real-world RL applications. The topics to be covered include model-based RL, which aims to address the data inefficiency and environmental uncertainty problems commonly encountered in deep RL, and offline RL techniques wherein we could only rely on the available dataset to learn policies when data collection is expensive. To shed light on interpreting agent behaviors, I will introduce works that combine attention with RL. Finally, manually designing tasks and tuning rewards to training policies can be exhausting, I will cover some works that focus on open-endedness RL that may train the desired policies and at the same time relieve us from the laborious task design.","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"From super-human performance in games to ultra-efficient automatic chip layout design, we have witnessed the rise of deep reinforcement learning (RL) and its transition from proof-of-concept tasks to beneficial real world applications. In this presentation, I will show some research topics that are critical for us to further advance toward more and more real-world RL applications. The topics to be covered include model-based RL, which aims to address the data inefficiency and environmental uncertainty problems commonly encountered in deep RL, and offline RL techniques wherein we could only rely on the available dataset to learn policies when data collection is expensive. To shed light on interpreting agent behaviors, I will introduce works that combine attention with RL. Finally, manually designing tasks and tuning rewards to training policies can be exhausting, I will cover some works that focus on open-endedness RL that may train the desired policies and at the same time relieve us from the laborious task design.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"1","bibliographic_titles":[{"bibliographic_title":"研究報告オーディオビジュアル複合情報処理（AVM）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2022-02-18","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"9","bibliographicVolumeNumber":"2022-AVM-116"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":216508,"links":{}}