ML-Promise
收藏资源简介:
ML-Promise是一个多语言数据集,旨在促进对环境、社会和治理(ESG)报告中企业承诺的深入验证。该数据集包含英语、法语、中文、日语和韩语,涵盖了来自英国、法国、台湾、日本和韩国的ESG报告。数据集的创建过程包括承诺识别、支持证据链接、承诺与证据清晰度评估以及验证时间推断。ML-Promise的应用领域主要集中在评估企业承诺的真实性和透明度,特别是在防止绿色清洗(greenwashing)方面。
ML-Promise is a multilingual dataset developed to facilitate in-depth validation of corporate commitments disclosed in environmental, social, and governance (ESG) reports. It encompasses ESG reports in five languages—English, French, Chinese, Japanese, and Korean—sourced from the United Kingdom, France, China's Taiwan, Japan, and the Republic of Korea. The construction of ML-Promise involves four key procedures: commitment identification, supporting evidence linking, evaluation of the clarity of both commitments and their associated supporting evidence, and inference of validation timelines. The primary application domains of ML-Promise center on assessing the authenticity and transparency of corporate commitments, with a particular focus on preventing greenwashing.

- 1ML-Promise: A Multilingual Dataset for Corporate Promise Verification筑波大学图书馆、信息与媒体科学研究所,筑波大学信息学院知识与图书馆科学学院,3DS Outscale,法国,新韩证券公司,韩国,国立台北大学信息管理研究生院,日本AIST · 2024年



