CausalMan
收藏资源简介:
CausalMan是一款基于物理的模拟器,由博世人工智能中心、达姆施塔特工业大学计算机科学系和黑森州人工智能中心共同开发,用于生成具有大规模因果关系的模拟数据。该数据集包含两个衍生数据集,模拟了现实世界生产线中复杂的线性与非线性的因果关系,并具有挑战性的预测行为,例如离散模式变化。数据集的设计满足了在制造领域中评价因果算法的需要,提供了观测数据和干预数据,有助于研究因果关系在不同规模下的行为。
CausalMan is a physics-based simulator co-developed by the Bosch Artificial Intelligence Center, the Department of Computer Science of Technische Universität Darmstadt, and the Hessian Center for Artificial Intelligence, dedicated to generating simulated data with large-scale causal relationships. This dataset includes two derived datasets that simulate complex linear and nonlinear causal relationships in real-world production lines, as well as challenging predictive behaviors such as discrete mode shifts. Designed to meet the needs of evaluating causal algorithms in the manufacturing domain, the dataset provides both observational and interventional data to facilitate research on the behavior of causal relationships across different scales.

- 1CausalMan: A physics-based simulator for large-scale causality博世人工智能中心,德国Renningen;计算机科学系,德国达姆施塔特工业大学;黑森州人工智能中心,德国 · 2025年



