[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85746-en":3,"doc-seo-85746-105":29,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85746,5909877438554,"Maeve","https://ap-avatar.wpscdn.com/avatar/5600025385ad2bf12a7?_k=1778553567797529272",8,"Research & Report","ShapKO Shapley-Adaptive Modality Knockout for Robust Multimodal Learning","Multimodal medical models often suffer when inputs are missing, and even with complete inputs they can become dominated by a single highly predictive modality during training. This over-optimization undertrains complementary sources, causing robustness collapse under partial modality availability. ShapKO (Shapley-Adaptive Modality Knockout) learns dynamic, modality-specific knockout probabilities from validation utility. Periodic subset evaluation estimates modality importance via Shapley values and suppresses dominant modalities more frequently, improving complementary representations. Experiments on three clinical tasks show consistent gains and interpretable masking behavior without architectural changes.","arXiv :2607 .09884v1 [ cs .CV] 10 Jul 2026  \nShapKO: Shapley-Adaptive Modality Knockout for Robust Multimodal Learning  \nNusrat Binta Nizam 1 ,2 ,3 , Fengbei Liu 1 ,3 , Sunwoo Kwak 1 ,2 ,3 , Minh Nguyen 1 ,2 ,3 ,  \nRuining Deng3 , and Mert R. Sabuncu 1 ,2 ,3  \n1 Cornell University, New York, NY, USA  \n2 Cornell Tech, New York, NY, USA  \n3 Weill Cornell Medicine, New York, NY, USA  \n{nn284,fl453,sk3355,bn244,[msabuncu}@cornell.edu](msabuncu}@cornell.edu) , [rud4004@med.cornell.edu](rud4004@med.cornell.edu)  \nAbstract. Multimodal medical models often degrade when inputs are missing, a common scenario in real-world clinical workflows. Separately, even when all modalities are present, modality dominance is observed during training, where optimization over-relies on a highly predictive modality and undertrains complementary sources, resulting in poor robustness under partial availability. While training-time modality knockout improves missing-modality robustness, existing approaches use static masking rates that cannot adapt to evolving modality utility during training. We introduce ShapKO (Shapley-Adaptive Modality Knockout), a dynamic training strategy that learns modality-specific knockout probabilities based on validation utility. ShapKO periodically evaluates performance across modality subsets, estimates modality importance via Shapley values, and updates masking probabilities to suppress dominant modalities more frequently. This adaptive process promotes complementary representations, while requiring no architectural modifications. We evaluate ShapKO on three datasets covering multitask clinical classification, survival prediction, and cancer detection. ShapKO consistently improves performance under modality absence and yields interpretable trajectories of learned masking behavior. Code is available at:  \n[https://github.com/sumona00/ShapKO](https://github.com/sumona00/ShapKO).  \nKeywords: Robust Multimodal learning · missing modalities · Shapley value · Knockout  \n1 Introduction  \nMultimodal learning aims to improve clinical decision support by combining complementary evidence from heterogeneous sources such as imaging, physiological signals, laboratory measurements, and structured clinical variables. In practice, modalities are often missing or corrupted due to cost and triage, conditional ordering, acquisition failures, and site-specific protocols [25] . Consequently, models trained under the assumption that all modalities are present at inference can degrade substantially under realistic missingness [14] .  \n2 NB. Nizam et al.  \nA key contributor is modality dominance: when one modality is consistently most predictive during training, optimization over-relies on it while undertraining weaker but complementary modalities, leading to substantial performance degradation when the dominant modality is unavailable [6,27] . Addressing this imbalance requires maintaining performance across diverse modality-availability patterns without training separate subset-specific models [9] . Existing robustness strategies, including modality masking, knockout, and gradient balancing, attempt to mitigate this issue by simulating missingness or reweighting modality contributions during training [15,25 , 16 , 13] . However, these approaches typically rely on static heuristics or local training signals and do not explicitly optimize performance across the combinatorial space of modality subsets. As a result, they cannot adapt to evolving modality utility dynamics and may over-regularize weaker modalities or insufficiently suppress dominant ones.  \nMissing modalities are common in clinical data and can substantially degrade multimodal performance, motivating methods that operate under arbitrary availability patterns [25] . In medical imaging, prior work includes anysubset fusion in a shared latent space for inference from partial modality sets [5] and generative/shared-latent formulations that support prediction and modality completion [2,24] ","cbCaihfd8wmQn7Cz","https://ap.wps.com/l/cbCaihfd8wmQn7Cz","pdf",872441,1,10,"English","en",105,"# Introduction\n## Modality dominance and robustness challenges\n## Related work on masking and balancing\n## ShapKO approach and contributions","[{\"question\":\"What problem does ShapKO address in multimodal medical learning?\",\"answer\":\"ShapKO targets performance degradation when modalities are missing and the training-time issue of modality dominance, where optimization relies too heavily on one predictive modality and undertrains complementary ones.\"},{\"question\":\"How does ShapKO adapt modality knockout during training?\",\"answer\":\"ShapKO periodically evaluates performance across modality subsets, estimates each modality’s importance using Shapley values, and updates per-modality knockout probabilities to mask more influential modalities more often.\"},{\"question\":\"What results does ShapKO report across experiments?\",\"answer\":\"ShapKO is evaluated on multitask clinical classification, survival prediction, and cancer detection, consistently improving performance under modality absence and producing interpretable trajectories of learned masking behavior without architectural modifications.\"}]",1784205988,25,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"shapko-shapley-adaptive-modality-knockout-for-robust-multimodal-learning","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/shapko-shapley-adaptive-modality-knockout-for-robust-multimodal-learning/85746/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What problem does ShapKO address in multimodal medical learning?","Question",{"text":74,"@type":75},"ShapKO targets performance degradation when modalities are missing and the training-time issue of modality dominance, where optimization relies too heavily on one predictive modality and undertrains complementary ones.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does ShapKO adapt modality knockout during training?",{"text":79,"@type":75},"ShapKO periodically evaluates performance across modality subsets, estimates each modality’s importance using Shapley values, and updates per-modality knockout probabilities to mask more influential modalities more often.",{"name":81,"@type":72,"acceptedAnswer":82},"What results does ShapKO report across experiments?",{"text":83,"@type":75},"ShapKO is evaluated on multitask clinical classification, survival prediction, and cancer detection, consistently improving performance under modality absence and producing interpretable trajectories of learned masking behavior without architectural modifications.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,127,130,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":21,"slug":132},"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]