[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85377-en":3,"doc-seo-85377-105":29,"detail-sidebar-cat-0-en-105":91},{"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":20,"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},85377,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",8,"Research & Report","Evidence-Backed Video Question Answering","Current video large language models excel at question answering but provide textual responses without verifiable visual grounding, limiting trust and explainability when videos involve occlusions and non-rigid motion. Evidence-Backed Video Question Answering (E-VQA) requires models to output an answer together with precise spatio-temporal evidence, including temporal segments and dense tracked pixel-level segmentation masklets. ST-Evidence is presented as the first human-verified benchmark for both discriminative and generative pixel grounding, and extensive evaluations reveal a decoupling between QA accuracy and true visual perception.","[ cs .CV] 13 Jul 2026  \nEvidence-Backed Video Question Answering  \nShijie Wang 1 ,2 , Honglu Zhou 1 , Ziyang Wang 1 , Ran Xu 1 , Caiming Xiong 1 , Silvio Savarese 1 , Chen Sun2 , and Juan Carlos Niebles 1  \n1 Salesforce, Palo Alto, CA, USA  \n2 Brown University, Providence, RI, USA  \nAbstract. Current Video Large Language Models (Video LLMs) excel in question answering (QA) but largely operate as black boxes, providing textual answers without verifiable visual grounding. Existing explainability efforts rely on textual rationales or sparse bounding boxes, which struggle to capture complex video dynamics such as occlusions and nonrigid deformations. We propose Evidence-Backed Video Question Answering (E-VQA), a novel task requiring models to jointly output a semantic answer and precise spatio-temporal evidence: temporal segments and dense, tracked object segmentation masklets. To support this, we introduce ST-Evidence, the first human-verified benchmark for both discriminative and generative pixel-level grounding. Evaluations of state-of-the-art models reveal a critical decoupling between QA accuracy and true visual perception that scaling alone fails to bridge.  \nTo address this, we develop scalable, automated generation pipelines to create ST-Evidence-Instruct, a 160k-scale dataset bridging highlevel reasoning with fine-grained grounding. Fine-tuning grounded Video LLMs on this data yields substantial gains over the corresponding sizematched UniPixel baselines (e.g., +27.2 t-mean and +13.8 J &F on a 7B model), establishing a robust baseline for explainable, evidencebacked video understanding. Code and data are available at [https:](https:)  \n2 S. Wang et al.  \nWhy did the man fall down?  \nThe boy hits the ball and hurts the man.  \n[[7 .5s, 9 . 8s]]  \nAnswer  \nTemporal Evidence  \nSpatial Evidence  \nFig. 1: E-VQA: Evidence-Backed Video Question Answering. Models provide textual answers to video questions while grounding their reasoning in spatio-temporal evidence, including relevant temporal video segments and densely tracked segmentation masks that highlight the spatio-temporal visual regions supporting the answer.  \nTo bridge the gap between semantic reasoning and visual perception, we argue that effective video understanding requires evidence grounding: the ability to explicitly justify an answer with its corresponding spatio-temporal visual evidence. Existing grounded video reasoning work [8, 30, 52] predominantly focuses on answer grounding (identifying the object that constitutes the answer), often relying on spatially sparse bounding boxes over temporally limited keyframes. However, such sparse grounding lacks the precision to capture complex video dynamics, such as severe occlusions, continuous state changes, non-rigid deformations (e.g., liquids, wires), or identity preservation through crossovers (e.g. , tracking a specific cup in a shell game) . We argue that to unambiguously verify that a model has perceived the correct visual cues, it must perform dense, pixel-level spatio-temporal tracking as a formal evidence trail.  \nTo formalize this task, we introduce Evidence-Backed Video Question Answering (E-VQA) (Fig. 1) . Given a video and a question, a model must generate a triplet: (i) the semantic answer, (ii) the supporting temporal evidence (relevant video segments), and (iii) the supporting spatial evidence represented as dense, tracked spatio-temporal segmentation masks (masklets) . By requiring these components, E-VQA explicitly couples high-level reasoning with low-level grounding. This formulation forces a model to not only solve the linguistic task but also to provide a verifiable, pixel-level justification of its perception.  \nAs no existing datasets support this setting, we introduce ST-Evidence, the first benchmark specifically designed for E-VQA. ST-Evidence is constructed through a rigorous three-stage semi-automatic pipeline: (1) Sample: Select suitable video–question–answer pairs from existing Video QA dat","cbCaidmBrksShyMm","https://ap.wps.com/l/cbCaidmBrksShyMm","pdf",16748372,1,19,"English","en",105,"# Evidence-Backed Video Question Answering (E-VQA)\n## ST-Evidence: Human-Verified Benchmark\n## Benchmark Variants: ST-Evidence-Gen and ST-Evidence-MCQ\n## Evaluation of Video LLMs and Findings","[{\"question\":\"What problem does E-VQA address in current video LLM question answering?\",\"answer\":\"E-VQA targets the lack of verifiable visual grounding in existing video QA models, where answers are not supported by precise spatio-temporal evidence in the video.\"},{\"question\":\"What evidence outputs does an E-VQA model generate?\",\"answer\":\"An E-VQA model generates a triplet: the semantic answer, the supporting temporal evidence (relevant video segments), and dense tracked spatial evidence as spatio-temporal segmentation masklets.\"},{\"question\":\"How is ST-Evidence constructed and verified?\",\"answer\":\"ST-Evidence uses a three-stage semi-automatic pipeline: sampling suitable video-question-answer pairs, human annotation of dense spatio-temporal evidence with rejection of low-quality samples, and independent human verification with re-annotation when needed.\"}]",1784202974,48,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":27},"evidence-backed-video-question-answering","",{"@graph":35,"@context":85},[36,53,68],{"@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/evidence-backed-video-question-answering/85377/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does E-VQA address in current video LLM question answering?","Question",{"text":75,"@type":76},"E-VQA targets the lack of verifiable visual grounding in existing video QA models, where answers are not supported by precise spatio-temporal evidence in the video.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What evidence outputs does an E-VQA model generate?",{"text":80,"@type":76},"An E-VQA model generates a triplet: the semantic answer, the supporting temporal evidence (relevant video segments), and dense tracked spatial evidence as spatio-temporal segmentation masklets.",{"name":82,"@type":73,"acceptedAnswer":83},"How is ST-Evidence constructed and verified?",{"text":84,"@type":76},"ST-Evidence uses a three-stage semi-automatic pipeline: sampling suitable video-question-answer pairs, human annotation of dense spatio-temporal evidence with rejection of low-quality samples, and independent human verification with re-annotation when needed.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]