Earnings-21

收益21数据集 (也称为earnings21) 是一个39小时的收益呼叫语料库，其中包含来自九个不同金融部门的实体密集语音。该语料库旨在在野外对自动语音识别 (ASR) 系统进行基准测试，并特别关注命名实体识别 (NER)。

The Earnings21 Dataset (also referred to as earnings21) is a 39-hour corpus of earnings calls, containing entity-dense speech from nine distinct financial sectors. This corpus is designed to benchmark automatic speech recognition (ASR) systems in the wild, with a particular focus on named entity recognition (NER).

搜集汇总

背景与挑战

背景概述

Earnings-21是一个39小时的收益呼叫语料库，涵盖九个金融部门的实体密集语音，主要用于自动语音识别(ASR)系统和命名实体识别(NER)的基准测试。该数据集由约翰霍普金斯大学于2021年发布，相关资源可通过提供的链接获取。