{"id":1185784,"date":"2026-09-10T08:59:02","date_gmt":"2026-09-10T15:59:02","guid":{"rendered":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/redeck-step-level-render-grounded-refinement-for-document-to-slide-generation\/"},"modified":"2026-09-30T15:09:48","modified_gmt":"2026-09-30T22:09:48","slug":"redeck-step-level-render-grounded-refinement-for-document-to-slide-generation","status":"publish","type":"msr-research-item","link":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/redeck-step-level-render-grounded-refinement-for-document-to-slide-generation\/","title":{"rendered":"ReDeck: Step-Level Render-Grounded Refinement for Document-to-Slide Generation"},"content":{"rendered":"\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Document-to-slide generation is challenging because slides are dense editable artifacts that require both faithful content selection and precise spatial layout. Recent slide agents adopt iterative reflection, but typically follow a monolithic&#8221;one version, one feedback&#8221;loop: a slide or deck is rewritten, rendered afterward, and critiqued only at the turn boundary. This delayed feedback makes local failures such as overflow, overlap, clipping, and off-canvas placement difficult to attribute and repair. We propose ReDeck, a step-level render-grounded refinement framework that decomposes slide revision into atomic edit actions and returns renderer-derived observations after each step, turning refinement into&#8221;one edit, one observation.&#8221;To balance local repair with global quality, ReDeck uses multi-granular feedback: step-level render feedback for spatial errors, a turn-level adaptive critic for semantic and design guidance, and a submission-level gate for hard layout validation. We further introduce DeckQuiz, a benchmark that decouples content fidelity, spatial correctness, and design quality. Across GPT-5.4, Claude-4.6, and Gemini-3.1, ReDeck consistently outperforms existing slide-generation agents, and ablations confirm that feedback timing and granularity are critical for reliable slide refinement.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Document-to-slide generation is challenging because slides are dense editable artifacts that require both faithful content selection and precise spatial layout. Recent slide agents adopt iterative reflection, but typically follow a monolithic&#8221;one version, one feedback&#8221;loop: a slide or deck is rewritten, rendered afterward, and critiqued only at the turn boundary. This delayed feedback makes local failures [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"text","value":"Mu-Zhao Tian","user_id":0},{"type":"text","value":"Ze-Zi Zeng","user_id":0},{"type":"user_nicename","value":"Yifan Yang","user_id":"41539"},{"type":"text","value":"Xin Gao","user_id":0},{"type":"text","value":"Yan Li","user_id":0},{"type":"text","value":"Zi-Su Huang","user_id":0},{"type":"text","value":"Xiaohua Wang","user_id":0},{"type":"text","value":"Changze Lv","user_id":0},{"type":"text","value":"Ming-Hung Cheng","user_id":0},{"type":"user_nicename","value":"Bei Liu","user_id":"38889"},{"type":"user_nicename","value":"Kai Qiu","user_id":"38988"},{"type":"user_nicename","value":"Qi Dai","user_id":"36689"},{"type":"user_nicename","value":"Dong Chen","user_id":"31661"},{"type":"user_nicename","value":"Yue Dong","user_id":"35060"},{"type":"text","value":"Xiao-Qing Zheng","user_id":0},{"type":"text","value":"Ji Li","user_id":0},{"type":"user_nicename","value":"Chong Luo","user_id":"31450"}],"msr_publishername":"","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"arXiv","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"","msr_doi":"","msr_arxiv_id":"2609.00194","msr_mag_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_release_tracker_id":"","msr_highlight_type":"","msr_date_display_format":"","msr_main_download_label":"","msr_external_link_label":"","msr_doi_label":"","msr_published_date":"2026-08-31","msr_startdate":"","msr_presentation_date":"","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_year":2026,"msr_month":8,"msr_day":31,"msr_microsoftintellectualproperty":false,"msr_pub_id":"0c4f696d1c3892ae9d7fc9140834ed57bc186acb","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2609.00194","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_original_fields_of_study":[],"msr_s2_paper_id":"0c4f696d1c3892ae9d7fc9140834ed57bc186acb","msr_s2_pdf_url":"","msr_citation_count_updated":"","msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":62,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[{"provider":"s2","id":"0c4f696d1c3892ae9d7fc9140834ed57bc186acb"},{"provider":"arxiv","id":"2609.00194"},{"provider":"corpusid","id":"291543809"}],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13556,13554],"msr-publication-type":[270373],"msr-publisher":[],"msr-publication-cta":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[256039,246691,255217],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-1185784","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-research-area-human-computer-interaction","msr-locale-en_us","msr-field-of-study-artifcial-intelligence","msr-field-of-study-computer-science","msr-field-of-study-human-computer-interaction-4"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2026-08-31","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"arXiv","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":0,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2609.00194","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"0c4f696d1c3892ae9d7fc9140834ed57bc186acb","msr_influential_citations":0,"msr_reference_count":62,"msr_arxiv_id":"2609.00194","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"text","value":"Mu-Zhao Tian","user_id":0,"rest_url":false},{"type":"text","value":"Ze-Zi Zeng","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Yifan Yang","user_id":41539,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Yifan Yang"},{"type":"text","value":"Xin Gao","user_id":0,"rest_url":false},{"type":"text","value":"Yan Li","user_id":0,"rest_url":false},{"type":"text","value":"Zi-Su Huang","user_id":0,"rest_url":false},{"type":"text","value":"Xiaohua Wang","user_id":0,"rest_url":false},{"type":"text","value":"Changze Lv","user_id":0,"rest_url":false},{"type":"text","value":"Ming-Hung Cheng","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Bei Liu","user_id":38889,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Bei Liu"},{"type":"user_nicename","value":"Kai Qiu","user_id":38988,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Kai Qiu"},{"type":"user_nicename","value":"Qi Dai","user_id":36689,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Qi Dai"},{"type":"user_nicename","value":"Dong Chen","user_id":31661,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Dong Chen"},{"type":"user_nicename","value":"Yue Dong","user_id":35060,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Yue Dong"},{"type":"text","value":"Xiao-Qing Zheng","user_id":0,"rest_url":false},{"type":"text","value":"Ji Li","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Chong Luo","user_id":31450,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Chong Luo"}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[],"msr_project":[],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"misc","related_content":[],"_links":{"self":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1185784","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":2,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1185784\/revisions"}],"predecessor-version":[{"id":1188040,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1185784\/revisions\/1188040"}],"wp:attachment":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1185784"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=1185784"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1185784"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=1185784"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=1185784"},{"taxonomy":"msr-publication-cta","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-cta?post=1185784"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=1185784"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1185784"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1185784"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=1185784"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=1185784"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=1185784"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1185784"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=1185784"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}