QUASAR-T (QUestion Answering by Search And Reading – Trivia)

QUASAR-T 是一个大规模数据集，旨在评估旨在理解自然语言查询并从大型文本语料库中提取其答案的系统。它由 43,013 个开放域琐事问题及其从各种互联网资源获得的答案组成。 ClueWeb09 作为提取这些答案的背景语料库。这些问题的答案是自由形式的文本，尽管大多数是名词短语。

QUASAR-T is a large-scale dataset designed to evaluate systems that aim to understand natural language queries and extract answers from large text corpora. It comprises 43,013 open-domain trivia questions along with their answers obtained from various internet resources. ClueWeb09 serves as the background corpus for extracting these answers. The answers to these questions are free-form text, though most are noun phrases.

搜集汇总

背景与挑战

背景概述

QUASAR-T是一个大规模开放域问答数据集，包含43,013个琐事问题及其从互联网资源获取的答案，主要用于评估系统从文本语料库中理解查询并提取答案的能力。该数据集以ClueWeb09为背景语料库，答案形式多为自由文本的名词短语。