[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-292335-105":59,"doc-detail-292335-en":131},{"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":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","bv-frd-a-multimodal-vietnamese-english-food-review-video-dataset-abstract","BV-FRD: A Multimodal Vietnamese-English Food Review Video Dataset - Abstract","","A short video summarization framework is presented to generate brief descriptions that help viewers grasp key information quickly. Despite progress in multimodal research, a shortage of food-review datasets—especially Vietnamese—persists. This work introduces BV-FRD, a multimodal Vietnamese-English dataset for Vietnamese food review videos, covering diverse dishes, prices, and locations, with processed scripts and annotated descriptions produced via an LLM-driven multi-stage pipeline plus human collaboration.",{"@graph":69,"@context":123},[70,84,106],{"@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/bv-frd-a-multimodal-vietnamese-english-food-review-video-dataset-abstract/292335/",{"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/bv-frd-a-multimodal-vietnamese-english-food-review-video-dataset-abstract/292335.png","ImageObject",300,407,{"name":92,"@type":93},"Felix Montgomery","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-30","2026-09-17",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",14,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What problem does BV-FRD address?","Question",{"text":113,"@type":114},"BV-FRD addresses the lack of Vietnamese-focused food review multimodal datasets, especially for Vietnamese video reviews that pair Vietnamese content with English descriptions.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"How is the BV-FRD dataset generated and annotated?",{"text":118,"@type":114},"It uses a multi-stage pipeline with several LLMs, and human collaboration verifies outputs. Each video includes processed scripts and annotated Vietnamese-English descriptions.",{"name":120,"@type":111,"acceptedAnswer":121},"What do the baseline experiments suggest about the dataset?",{"text":122,"@type":114},"Baseline experiments show moderate performance, indicating the task is challenging. DeepSeek performs best across three of four metrics in both Vietnamese and English, though generated descriptions remain only relatively similar to ground truth.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},292335,1789650214,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":140,"language":141,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":67,"update_tm":130,"read_time":36,"preview_page_count":4},549768064778,"https://ap-avatar.wpscdn.com/davatar_6f874abed73319feea01a86fa6f0fab8","BV-FRD: A Multimodal Vietnamese-English Food Review Video  \nDescription Generation  \nBao Pham-Thai1,2 , Vy Do Le Khanh1,2 , Huy Quoc To1,2  \n1University of Information Technology, Ho Chi Minh city, Vietnam  \n2Vietnam National University, Ho Chi Minh city, Vietnam  \n{21520156, [23521826}@gm.uit.edu.vn](23521826}@gm.uit.edu.vn) , [huytq@uit.edu.vn](huytq@uit.edu.vn)  \nAbstract  \nThe short video summarization provides brief descriptions ofthe main content, enabling viewers to quickly understand the key information.  \nAlthough the field has seen significant progress, there remains a shortage of datasets for food reviews, particularly in Vietnamese. In this paper, we introduce a novel multimodal Vietnamese  \nEnglish dataset focused on Vietnamese food review videos called BV-FRD. Our dataset includes a wide range of food, prices and restaurant locations. Each video includes processed scripts and annotated Vietnamese-English descriptions, generated through a multi-stage pipeline using several LLMs with human collaboration. Baseline experiments show moderate performance, indicating that the dataset is challenging and has strong potential for practical applications. In our experiments, DeepSeek outperforms other models in Vietnamese and English across three of four evaluation metrics. In Vietnamese, Phi-4 achieves the highest BERTS score, with a value of 0.64 precision. In English, DeepSeek reaches the highest consistency in Uni-Eval, with a value of 0.78. Through our analysis and baseline experiments, we demonstrate that our dataset is valuable and challenging for multimodal food review description generation task. Our dataset is available through this link1.  \n1 Introduction  \nEvery day, millions engage with short videos related to food, shaping food trends, domestic tourism, and global perceptions of Vietnamese cuisine (Truong and Kim, 2023) . These videos on social networks attract the audience through their brevity, practicality, and high shareability (Violot et al., 2024 ; Zannettou et al., 2024) . In addition, it can affect consumers’ choices, identity construction, and cultural preservation.  \n1 [https://anonymous.4open.science/r/BV-FRD](https://anonymous.4open.science/r/BV-FRD)  \nMost of this content is open source, providing multi-modal datasets that combine speech, facial expressions, visuals, and sounds, offering a clearer view of real-world communication than textonly data (Baltrušaitis et al., 2018) . Open-source data promotes transparency, reproducibility, and supports research in artificial intelligence and human–computer interaction (Von Krogh et al., 2003 ; Willmes et al., 2014) .  \nIn recent years, Vietnamese natural language processing (NLP) has advanced significantly, with high-quality studies on word segmentation (Hai et al., 2025), document summarization (Le and Le, 2025), and social media text processing (Nguyen et al., 2023) . Pre-trained models like PhoBERT (Nguyen and Nguyen, 2020) have boosted performance across downstream tasks. Research on sentiment analysis and question-answering systems (Lê et al., 2020) further enriches the NLP landscape. Overall, these developments lay the foundation for advanced Vietnamese language models capable of contextual understanding, supporting applications from automated customer support to content creation.  \nAlthough a niche domain, food reviews strongly influence consumers, promote local cuisine, and support cultural exchange (Rini et al., 2024) . Research on short Vietnamese videos is limited, and open-source multimodal datasets combining images, audio, and text are lacking, which are crucial for conversational AI and multimodal understanding (Baltrušaitis et al., 2018) .  \nWe present the Vietnamese–English bilingual dataset BV-FRD for short food-related videos, containing transcripts, manually written summaries, and emotion labels to enrich Vietnamese language resources and support research on dialogue summarization, multimodal sentiment analysis, and contextual understa","cbCais8B5XIgD1fW","https://ap.wps.com/l/cbCais8B5XIgD1fW","pdf",21076777,16,"English","# Abstract\n# Introduction\n## Open-source multimodal data context\n## Vietnamese NLP foundations\n## Motivation for BV-FRD\n## Contributions and dataset details\n# Related Work\n## Multimodal era","[{\"question\":\"What problem does BV-FRD address?\",\"answer\":\"BV-FRD addresses the lack of Vietnamese-focused food review multimodal datasets, especially for Vietnamese video reviews that pair Vietnamese content with English descriptions.\"},{\"question\":\"How is the BV-FRD dataset generated and annotated?\",\"answer\":\"It uses a multi-stage pipeline with several LLMs, and human collaboration verifies outputs. Each video includes processed scripts and annotated Vietnamese-English descriptions.\"},{\"question\":\"What do the baseline experiments suggest about the dataset?\",\"answer\":\"Baseline experiments show moderate performance, indicating the task is challenging. DeepSeek performs best across three of four metrics in both Vietnamese and English, though generated descriptions remain only relatively similar to ground truth.\"}]","BV-FRD: A Multimodal Vietnamese-English Food Review Video Dataset - Abstract | PDF"]