将因果发现中的模糊关系转化为确定图结构，提升下游应用可靠性。
No More Maybe-Arrows: Resolving Causal Uncertainty by Breaking Symmetries
- 通过状态级表示扩展离散变量，缩小搜索空间
- 联合优化后生成的图能保留真实因果关系且无环
- 适合需要确定因果结构的研究者使用
近年来因果发现研究普遍采用部分祖先图（PAG）建模，因观测数据仅能将真实因果有向无环图（DAG）约束至马尔可夫等价类内，导致因果关系仍存在不确定性，限制了其在多数下游任务中的应用。本文提出新框架 CausalSAGE，将 PAG 转化为 DAG，同时尊重潜在因果关系。该框架将离散变量扩展为状态级表示，利用结构知识与软先验约束搜索空间，并采用统一可微目标进行联合优化。最终通过聚合优化结构并必要时强制无环性，获得 DAG。实验表明，所得 DAG 既能保持底层因果关系，又具备高效性。
原文摘要 · Abstract (English)
The recent works on causal discovery have followed a similar trend of learning partial ancestral graphs (PAGs) since observational data constrain the true causal directed acyclic graph (DAG) only up to a Markov equivalence class. This limits their application in the majority of downstream tasks, as uncertainty in causal relations remains unresolved. We propose a new refinement framework, CausalSAGE, for converting PAGs to DAGs while respecting the underlying causal relations. The framework expands discrete variables into state-level representations, constrains the search space using structural knowledge and soft priors, and applies a unified differentiable objective for joint optimization. The final DAG is obtained by aggregating the optimized structures and enforcing acyclicity when necessary. Our experimental evaluations show that the obtained DAGs preserve the underlying causal relations while also being efficient to obtain.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。