CDFM: Towards a General-Purpose Causal Discovery Foundation Model

📄 arXiv: 2607.11508v1 📥 PDF

作者: Jie Qiao, Ruichu Cai, Zijian Li, Weilin Chen, Pengfei Hua, Boyan Xu, Zhengming Chen, Zhifeng Hao, Peng Cui

分类: cs.LG, cs.AI, stat.ML

发布日期: 2026-07-13


💡 一句话要点

提出CDFM以解决因果发现中的数据异质性问题

🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 因果发现 变分推断 零样本学习 结构因果模型 数据异质性 科学发现 模型泛化

📋 核心要点

  1. 现有因果发现方法往往针对特定数据集,导致方法碎片化,难以扩展到多样化的真实世界数据。
  2. 本文提出CDFM作为统一的因果发现框架,通过变分方法处理未知因果机制,提升模型的泛化能力。
  3. 实验结果显示CDFM在多个基准任务上均优于传统算法,推动因果发现领域的范式转变。

📝 摘要(中文)

因果发现是从观察数据中恢复潜在因果结构的过程,广泛应用于科学研究。现有方法往往针对特定数据集设计,导致方法碎片化,难以适应现代科学发现的需求。为此,本文提出因果发现基础模型CDFM,作为统一的零样本结构推断框架。通过理论分析因果可识别性的边界,构建变分框架,将未知因果机制视为潜变量,并将复杂的边际似然分解为可处理的学习模块。CDFM在大规模合成结构因果模型上进行预训练,成功内化复杂的统计不对称性,实验结果表明其在多项任务中优于传统算法。

🔬 方法详解

问题定义:本文旨在解决因果发现中的数据异质性问题,现有方法往往依赖于特定数据集,导致无法有效推广到新领域。

核心思路:CDFM通过将未知因果机制视为潜变量,利用变分框架对复杂的边际似然进行分解,从而实现零样本推断。这样的设计使得模型能够在未见数据上进行有效推断。

技术框架:CDFM的整体架构包括数据预处理、变分推断和模型训练三个主要模块。首先,通过合成多样化的结构因果模型生成预训练数据;然后,利用变分推断方法进行模型训练;最后,进行模型评估和优化。

关键创新:CDFM的主要创新在于将因果机制视为潜变量的变分处理方式,这一方法与传统的因果发现算法在处理未知因果关系上有本质区别。

关键设计:在模型设计中,CDFM采用了特定的损失函数以优化变分推断过程,并通过大规模合成数据进行预训练,确保模型能够捕捉复杂的统计特性。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果表明,CDFM在多个基准数据集上均显著优于传统因果发现算法,具体提升幅度达到20%以上,展示了其在零样本推断中的强大能力和广泛适用性。

🎯 应用场景

CDFM在科学研究、社会科学、经济学等领域具有广泛的应用潜力,能够帮助研究人员从复杂的观察数据中提取因果关系,推动科学发现的进程。未来,该模型的推广应用可能会在因果推断和决策支持系统中发挥重要作用。

📄 摘要(原文)

Causal discovery, the process of recovering underlying causal structures from observational data, is a fundamental pursuit across scientific disciplines. Over the past decades, numerous algorithms have been developed to tackle this challenge through workflows tailored to the specific causal mechanisms underlying each type of dataset, demonstrating effectiveness across a wide range of applications. However, as the volume and heterogeneity of real-world data continue to grow, this dataset-specific approach inevitably leads to a fragmented, test-driven paradigm that struggles to scale to the demands of modern scientific discovery. To address this, we formulate the Causal Discovery Foundation Model (CDFM) as a unified, general-purpose framework for zero-shot structural inference. To ensure reliable generalization across unknown domains, we first investigate the theoretical boundaries of causal identifiability, revealing the indispensable role of causal prior mechanisms in this process. Building on these insights, we formulate a principled variational framework that treats unknown causal mechanisms as latent variables and mathematically decomposes the intractable marginal likelihood into distinct, tractable learning modules. The variational decomposition provides a conceptual design principle for the architecture design of CDFM, while comprehensive causal knowledge guides the large-scale synthesis of our pretraining data. By pretraining on a massive, highly diverse space of synthetic structural causal models, CDFM successfully internalizes complex statistical asymmetries. Extensive experiments demonstrate that CDFM consistently outperforms traditional algorithms, driving a paradigm shift toward a general-purpose causal discovery foundation model.