MultiTEND
收藏资源简介:
MultiTEND是由香港理工大学等机构提出的首个大型多语言自然语言至NoSQL查询生成基准数据集,包含英语、德语、法语、俄语、日语和普通话六种语言。该数据集通过半自动化管道将单语言数据集扩展为多语言版本,包括数据库字段、自然语言查询(NLQ)和NoSQL查询的翻译。数据集涵盖了101,789个NLQ和20,351个相应的查询,旨在解决多语言环境下自然语言至NoSQL查询生成的结构性和词汇性挑战。
MultiTEND is the first large-scale multilingual benchmark dataset for natural language-to-NoSQL query generation, proposed by The Hong Kong Polytechnic University and other institutions. It encompasses six languages: English, German, French, Russian, Japanese, and Mandarin Chinese. This dataset expands monolingual datasets into multilingual variants through a semi-automated pipeline, covering the translation of database fields, natural language queries (NLQs), and NoSQL queries. It comprises 101,789 NLQs and 20,351 corresponding NoSQL queries, and is designed to address the structural and lexical challenges in natural language-to-NoSQL query generation within multilingual environments.

- 1MultiTEND: A Multilingual Benchmark for Natural Language to NoSQL Query Translation香港理工大学 · 2025年



