[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-273373-105":59,"doc-detail-273373-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","geoexplain-multimodal-reasoning-based-on-hierarchy-of-visual-information-in-street-view-abstract","GeoExplain: Multimodal Reasoning based on Hierarchy of Visual Information in Street View - Abstract","","Multimodal reasoning integrates and infers information across different data modalities and remains an active research focus. Existing evaluation tasks show limitations, especially in reasoning over hierarchical visual clues spanning local details and global context. To address this gap, GeoExplain is introduced as a dataset for explainable geolocalization, requiring both location prediction from street-view panoramas and detailed human-expert explanations. The dataset includes 40350 panorama-location-explanation tuples, and a multimodal multilevel method, SightSense, achieves strong performance. The dataset is available online.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/geoexplain-multimodal-reasoning-based-on-hierarchy-of-visual-information-in-street-view-abstract/273373/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/geoexplain-multimodal-reasoning-based-on-hierarchy-of-visual-information-in-street-view-abstract/273373.png","ImageObject",300,407,{"name":92,"@type":93},"Jiven","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-30","2026-09-15",true,{"@type":102,"interactionType":103,"userInteractionCount":29},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What is GeoExplain and what task does it evaluate?","Question",{"text":112,"@type":113},"GeoExplain evaluates explainable geolocalization. Given street-view panoramas, the task predicts the location and provides a detailed explanation of how the location is inferred.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How are explanations defined in GeoExplain?",{"text":117,"@type":113},"Each instance includes a set of panoramas, a street-level location, and human-expert explanations. The explanations describe how visual content cues lead to the predicted location.",{"name":119,"@type":110,"acceptedAnswer":120},"What is SightSense?",{"text":121,"@type":113},"SightSense is a multimodal multilevel reasoning method designed to make predictions and generate comprehensive explanations for GeoExplain.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},273373,1789471111,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":29,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":129,"read_time":144,"preview_page_count":4},1099513958607,"https://ap-avatar.wpscdn.com/avatar/100002390cf8733938c?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778829742770036399","arXiv : 2506 . 16633v 3 [cs .CL] 2 1 Aug 2026  \nGeoExplain: Multimodal Reasoning based on Hierarchy of Visual  \nInformation in Street View  \nFenghua Cheng  \n[fenghua.cheng@uq.edu.au](fenghua.cheng@uq.edu.au)[ ](fenghua.cheng@uq.edu.au)The University of Queensland Brisbane, Queensland, Australia  \nJinxiang Wang  \n[jinxiang.wang@uq.edu.au](jinxiang.wang@uq.edu.au)[ ](jinxiang.wang@uq.edu.au)The University of Queensland Brisbane, Queensland, Australia  \nSen Wang  \n[sen.wang@uq.edu.au](sen.wang@uq.edu.au)[ ](sen.wang@uq.edu.au)The University of Queensland Brisbane, Queensland, Australia  \nZi Huang  \n[helen.huang@uq.edu.au](helen.huang@uq.edu.au)[ ](helen.huang@uq.edu.au)The University of Queensland Brisbane, Queensland, Australia  \nXue Li  \n[xueli@eecs.uq.edu.au](xueli@eecs.uq.edu.au)[ ](xueli@eecs.uq.edu.au)The University of Queensland Brisbane, Queensland, Australia  \nCoordinates  \nLat=20.39319610595703, Lng=-101.19889068603516  \nElection season again. This time with people running for\"presidente municipal\" instead of \"alcalde provincial \"meaning we're probably not in Peru this round. Sun in the south and this kind of cityscape means we're probably in Mexico. I did not find Valle de Santiago and guessed near Mexico City.  \nPanoramas Explanations  \nFigure 1: A GeoExplain example: each sample consists of a set of panoramas, location and explanation on how to infer the location. The explanation progressively integrates fine-grained cues (administrative terminology and language) and global context (cityscape and solar direction) to infer the country and further narrow the prediction to a specific city.  \nAbstract  \nMultimodal reasoning is a process of understanding, integrating and inferring information across different data modalities. It has recently attracted surging academic attention. Although there are various tasks for evaluating multimodal reasoning ability, they still have limitations. Reasoning on hierarchical visual clues at different levels of granularity, i.e., local details and global context, is of little discussion, despite its frequent involvement in  \nhuman reasoning. To bridge the gap, we introduce a challenging dataset, namely GeoExplain, which evaluates explainable geolocalization. Given a street view image, the task is to predict its location and provide a detailed explanation. GeoExplain consists of 40350 panoramas-location-explanation tuples. Each instance contains a set of street-view panoramas, a location on street level, and human-expert explanations describing how the location can be inferred from the visual content of panoramas. Additionally, we present a multimodal and multilevel reasoning method, namely  \nSightSense which can make predictions and generate a comprehensive explanation. Our analysis and experiments demonstrate its outstanding performance in GeoExplain. The dataset can be accessed via: [https://anonymous.4open.science/r/GeoExplain-A954](https://anonymous.4open.science/r/GeoExplain-A954) .  \nCCS Concepts  \n• Computing methodologies → Natural language generation.  \nKeywords  \nMultimodal Reasoning, Multimodal Dataset, Image Geo-Localization, Hierarchical Visual Information  \n1 Introduction  \nHumans cannot only see the world, but more importantly comprehend the world through it. Observing, associating visual or textual knowledge, making interpretation, and even building a cognitive framework demonstrates a critical and widely used ability of human intelligence, known as multimodal reasoning. Therefore, developing more human-like multimodal reasoning ability is an essential step towards advancing Artificial General Intelligence (AGI) [18] .  \nIn recent years, there has been significant enthusiasm for researching on Large Language Model (LLM) [6, 13, 53, 66] driven by the rise of self-attention mechanism and Transformer architecture [67] . We have witnessed the remarkable performance of LLM on Natural Language Understanding [46, 54, 79] . The great progress prompted the AI community to embark on","cbCaicgbUvvRWj96","https://ap.wps.com/l/cbCaicgbUvvRWj96","pdf",13061845,11,"English","# Abstract\n# Keywords\n# 1 Introduction","[{\"question\":\"What is GeoExplain and what task does it evaluate?\",\"answer\":\"GeoExplain evaluates explainable geolocalization. Given street-view panoramas, the task predicts the location and provides a detailed explanation of how the location is inferred.\"},{\"question\":\"How are explanations defined in GeoExplain?\",\"answer\":\"Each instance includes a set of panoramas, a street-level location, and human-expert explanations. The explanations describe how visual content cues lead to the predicted location.\"},{\"question\":\"What is SightSense?\",\"answer\":\"SightSense is a multimodal multilevel reasoning method designed to make predictions and generate comprehensive explanations for GeoExplain.\"}]","GeoExplain: Multimodal Reasoning based on Hierarchy of Visual Information in Street View - Abstract | PDF",28]