{"id":1185786,"date":"2026-09-10T08:59:02","date_gmt":"2026-09-10T15:59:02","guid":{"rendered":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/sparse-concept-attribution-for-histomorphological-hypothesis-generation-from-whole-slide-classifiers\/"},"modified":"2026-09-30T15:12:13","modified_gmt":"2026-09-30T22:12:13","slug":"sparse-concept-attribution-for-histomorphological-hypothesis-generation-from-whole-slide-classifiers","status":"publish","type":"msr-research-item","link":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/sparse-concept-attribution-for-histomorphological-hypothesis-generation-from-whole-slide-classifiers\/","title":{"rendered":"Sparse concept attribution for histomorphological hypothesis generation from whole-slide classifiers"},"content":{"rendered":"\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Histology images contain rich morphological information and can provide insights into pathological processes. However, deriving hypotheses relating morphological phenotypes to clinical attributes is bottlenecked by a manual image interpretation step. Here, we demonstrate that this process can be automated through interpretable deep learning. We present SCOPE, a method to interpret slide-level classifiers by combining pathology-specific vision&#8211;language models with sparse concept attribution onto a generalist histomorphological concept bank. To measure whether such explanations recover known morphology, we introduce MorphoRecoveryBench, a benchmark of seven tasks with pathologist-curated reference descriptions. On this benchmark, dense concept attribution is indistinguishable from a random baseline, whereas sparse attribution recovers substantial known morphology; decomposing the pooled slide embedding reaches similar explanation correctness at a fraction of the computational cost. Post-hoc interpretation of whole-slide classifiers can thus generate morphological hypotheses at scale, for expert validation.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Histology images contain rich morphological information and can provide insights into pathological processes. However, deriving hypotheses relating morphological phenotypes to clinical attributes is bottlenecked by a manual image interpretation step. Here, we demonstrate that this process can be automated through interpretable deep learning. We present SCOPE, a method to interpret slide-level classifiers by combining pathology-specific [&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":"user_nicename","value":"Tristan Lazard","user_id":"43910"},{"type":"user_nicename","value":"Kenza Bouzid","user_id":"43290"},{"type":"text","value":"Julius Hense","user_id":0},{"type":"user_nicename","value":"Shruthi Bannur","user_id":"39213"},{"type":"user_nicename","value":"Daniel Coelho de Castro","user_id":"39811"},{"type":"text","value":"Daniel Shao","user_id":0},{"type":"user_nicename","value":"Rajesh Jena","user_id":"36224"},{"type":"text","value":"Drew Williamson","user_id":0},{"type":"user_nicename","value":"Stephanie Hyland","user_id":"38458"}],"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":"2609.02985","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":"2026-09-02","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":9,"msr_day":2,"msr_microsoftintellectualproperty":false,"msr_pub_id":"2cb1c2f9dccf98bcd7f48756703daaae04223b0d","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2609.02985","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_original_fields_of_study":[],"msr_s2_paper_id":"2cb1c2f9dccf98bcd7f48756703daaae04223b0d","msr_s2_pdf_url":"","msr_citation_count_updated":"","msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":0,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[{"provider":"s2","id":"2cb1c2f9dccf98bcd7f48756703daaae04223b0d"},{"provider":"arxiv","id":"2609.02985"},{"provider":"corpusid","id":"291678341"}],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13562,13553],"msr-publication-type":[270373],"msr-publisher":[],"msr-publication-cta":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[256039,250972,248026,266442],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-1185786","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-computer-vision","msr-research-area-medical-health-genomics","msr-locale-en_us","msr-field-of-study-artifcial-intelligence","msr-field-of-study-biology","msr-field-of-study-engineering","msr-field-of-study-quantitative-methods"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2026-09-02","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\/2609.02985","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"2cb1c2f9dccf98bcd7f48756703daaae04223b0d","msr_influential_citations":0,"msr_reference_count":0,"msr_arxiv_id":"2609.02985","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"user_nicename","value":"Tristan Lazard","user_id":43910,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Tristan Lazard"},{"type":"user_nicename","value":"Kenza Bouzid","user_id":43290,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Kenza Bouzid"},{"type":"text","value":"Julius Hense","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Shruthi Bannur","user_id":39213,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Shruthi Bannur"},{"type":"user_nicename","value":"Daniel Coelho de Castro","user_id":39811,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Daniel Coelho de Castro"},{"type":"text","value":"Daniel Shao","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Rajesh Jena","user_id":36224,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Rajesh Jena"},{"type":"text","value":"Drew Williamson","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Stephanie Hyland","user_id":38458,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Stephanie Hyland"}],"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\/1185786","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\/1185786\/revisions"}],"predecessor-version":[{"id":1188042,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1185786\/revisions\/1188042"}],"wp:attachment":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1185786"}],"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=1185786"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1185786"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=1185786"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=1185786"},{"taxonomy":"msr-publication-cta","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-cta?post=1185786"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=1185786"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1185786"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1185786"},{"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=1185786"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=1185786"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=1185786"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1185786"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=1185786"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}