Open Causal Discovery Benchmark (OCDB)
收藏资源简介:
Open Causal Discovery Benchmark (OCDB) 是由华中科技大学开发的一个基于真实数据的综合性因果发现基准数据集。该数据集旨在通过包含多种类型的真实数据,全面评估因果发现算法的性能,以推动大型语言模型(LLMs)的解释性和可靠性。OCDB 数据集涵盖了广泛的复杂场景和多样化的数据源,支持因果结构和因果效应的差异评估,有助于选择更合适的因果发现方法,从而提高LLMs的解释性和可信度。该数据集的应用领域包括但不限于医疗和金融等高风险领域,旨在解决因果关系理解和决策支持的问题。
Open Causal Discovery Benchmark (OCDB) is a comprehensive real-world data-based causal discovery benchmark dataset developed by Huazhong University of Science and Technology. This dataset aims to comprehensively evaluate the performance of causal discovery algorithms by incorporating diverse types of real-world data, thereby advancing the interpretability and reliability of Large Language Models (LLMs). The OCDB dataset covers a wide range of complex scenarios and diverse data sources, supporting comparative evaluations of causal structures and causal effects, which assists in selecting more appropriate causal discovery methods and consequently enhancing the interpretability and credibility of LLMs. Its application fields include but are not limited to high-risk domains such as healthcare and finance, with the goal of addressing issues related to causal relationship understanding and decision support.

- 1OCDB: Revisiting Causal Discovery with a Comprehensive Benchmark and Evaluation Framework华中科技大学 · 2024年



