Zhuoyue Wan, Yuanfeng Song, Shuaimin Li, Chen Jason Zhang, Raymond Chi-Wing Wong
{"title":"DataVisT5:用于联合理解文本和数据可视化的预训练语言模型","authors":"Zhuoyue Wan, Yuanfeng Song, Shuaimin Li, Chen Jason Zhang, Raymond Chi-Wing Wong","doi":"arxiv-2408.07401","DOIUrl":null,"url":null,"abstract":"Data visualization (DV) is the fundamental and premise tool to improve the\nefficiency in conveying the insights behind the big data, which has been widely\naccepted in existing data-driven world. Task automation in DV, such as\nconverting natural language queries to visualizations (i.e., text-to-vis),\ngenerating explanations from visualizations (i.e., vis-to-text), answering\nDV-related questions in free form (i.e. FeVisQA), and explicating tabular data\n(i.e., table-to-text), is vital for advancing the field. Despite their\npotential, the application of pre-trained language models (PLMs) like T5 and\nBERT in DV has been limited by high costs and challenges in handling\ncross-modal information, leading to few studies on PLMs for DV. We introduce\n\\textbf{DataVisT5}, a novel PLM tailored for DV that enhances the T5\narchitecture through a hybrid objective pre-training and multi-task fine-tuning\nstrategy, integrating text and DV datasets to effectively interpret cross-modal\nsemantics. Extensive evaluations on public datasets show that DataVisT5\nconsistently outperforms current state-of-the-art models on various DV-related\ntasks. We anticipate that DataVisT5 will not only inspire further research on\nvertical PLMs but also expand the range of applications for PLMs.","PeriodicalId":501123,"journal":{"name":"arXiv - CS - Databases","volume":"440 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"DataVisT5: A Pre-trained Language Model for Jointly Understanding Text and Data Visualization\",\"authors\":\"Zhuoyue Wan, Yuanfeng Song, Shuaimin Li, Chen Jason Zhang, Raymond Chi-Wing Wong\",\"doi\":\"arxiv-2408.07401\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Data visualization (DV) is the fundamental and premise tool to improve the\\nefficiency in conveying the insights behind the big data, which has been widely\\naccepted in existing data-driven world. Task automation in DV, such as\\nconverting natural language queries to visualizations (i.e., text-to-vis),\\ngenerating explanations from visualizations (i.e., vis-to-text), answering\\nDV-related questions in free form (i.e. FeVisQA), and explicating tabular data\\n(i.e., table-to-text), is vital for advancing the field. Despite their\\npotential, the application of pre-trained language models (PLMs) like T5 and\\nBERT in DV has been limited by high costs and challenges in handling\\ncross-modal information, leading to few studies on PLMs for DV. We introduce\\n\\\\textbf{DataVisT5}, a novel PLM tailored for DV that enhances the T5\\narchitecture through a hybrid objective pre-training and multi-task fine-tuning\\nstrategy, integrating text and DV datasets to effectively interpret cross-modal\\nsemantics. Extensive evaluations on public datasets show that DataVisT5\\nconsistently outperforms current state-of-the-art models on various DV-related\\ntasks. We anticipate that DataVisT5 will not only inspire further research on\\nvertical PLMs but also expand the range of applications for PLMs.\",\"PeriodicalId\":501123,\"journal\":{\"name\":\"arXiv - CS - Databases\",\"volume\":\"440 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-14\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Databases\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.07401\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Databases","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.07401","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
数据可视化(Data Visualization,DV)是提高传达大数据背后见解效率的基础和前提工具,在现有的数据驱动世界中已被广泛接受。DV 中的任务自动化,如将自然语言查询转换为可视化(即文本到可视化)、从可视化生成解释(即可视化到文本)、以自由形式回答 DV 相关问题(即 FeVisQA)以及阐释表格数据(即表格到文本),对于推动该领域的发展至关重要。尽管预训练语言模型(PLMs)(如 T5 和 BERT)潜力巨大,但其在 DV 中的应用却因成本高昂和处理跨模态信息的挑战而受到限制,导致针对 DV 的预训练语言模型的研究寥寥无几。我们介绍了textbf{DataVisT5},这是一种为DV量身定制的新型PLM,它通过混合目标预训练和多任务微调策略增强了T5架构,整合了文本和DV数据集,从而有效地解释了跨模态语义。在公共数据集上进行的广泛评估表明,DataVisT5 在各种 DV 相关任务上的表现始终优于当前最先进的模型。我们预计,DataVisT5 不仅会激发对垂直 PLM 的进一步研究,而且还会扩大 PLM 的应用范围。
DataVisT5: A Pre-trained Language Model for Jointly Understanding Text and Data Visualization
Data visualization (DV) is the fundamental and premise tool to improve the
efficiency in conveying the insights behind the big data, which has been widely
accepted in existing data-driven world. Task automation in DV, such as
converting natural language queries to visualizations (i.e., text-to-vis),
generating explanations from visualizations (i.e., vis-to-text), answering
DV-related questions in free form (i.e. FeVisQA), and explicating tabular data
(i.e., table-to-text), is vital for advancing the field. Despite their
potential, the application of pre-trained language models (PLMs) like T5 and
BERT in DV has been limited by high costs and challenges in handling
cross-modal information, leading to few studies on PLMs for DV. We introduce
\textbf{DataVisT5}, a novel PLM tailored for DV that enhances the T5
architecture through a hybrid objective pre-training and multi-task fine-tuning
strategy, integrating text and DV datasets to effectively interpret cross-modal
semantics. Extensive evaluations on public datasets show that DataVisT5
consistently outperforms current state-of-the-art models on various DV-related
tasks. We anticipate that DataVisT5 will not only inspire further research on
vertical PLMs but also expand the range of applications for PLMs.