QFinder: A Framework for Quantity-centric Ranking

SIGIR '22: Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval(2022)

引用 2|浏览6
暂无评分
摘要
Quantities shape our understanding of measures and values, and they are an important means to communicate the properties of objects. Often, search queries contain numbers as retrieval units, e.g., "iPhone that costs less than 800 Euros''. Yet, modern search engines lack a proper understanding of numbers and units. In queries and documents, search engines handle them as normal keywords and therefore are ignorant of relative conditions between numbers, such as greater than or less than, or, more generally, the numerical proximity of quantities. In this work, we demonstrate QFinder, our quantity-centric framework for ranking search results for queries with quantity constraints. We also open-source our new ranking method as an Elasticsearch plug-in for future use. Our demo is available at: https://qfinder.ifi.uni-heidelberg.de/
更多
查看译文
关键词
Ranking functions, Quantities, Information Extraction
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要