Causal Inference Algorithm Driven by Denoising Diffusion Probabilistic Model

doi:10.3969/j.issn.1006-2475.2025.11.015

Abstract

Abstract: Abstract：A new causal inference algorithm is proposed to address the limitations of traditional methods， which fail to account for the impact of interventions on structural causal models and suffer from large inference bias and low stability when handling high-dimensional， complex causal data. This algorithm integrates a denoising diffusion probabilistic model. First， it analyzes the effects of different types of interventions on both the structural causal model and the diffusion model to improve algorithm interpretability. Next， the likelihood lower bound of the model under the causal Markov assumption is theoretically derived， and a diffusion sampling process that includes causal parents is constructed. Finally， by combining the variational structure of the diffusion-based causal model， the algorithm can sample from both original and intervention data and perform counterfactual estimation， thus simplifying model training complexity and enhancing robustness. In comparison with other algorithms， simulation results show that under different structural causal model assumptions， the proposed algorithm reduces the maximum mean discrepancy by 10%~41% and the mean squared error of counterfactual estimation by 9%~63%. Empirical results indicate a reduction of 11% in the maximum mean discrepancy and 5% in the mean squared error of counterfactual estimation. Experiments demonstrate that this algorithm effectively handles complex data structures and noise distributions， significantly improving the accuracy and stability of both sampling and counterfactual estimation.

Key words: Key words: structural causal model, neural network model, denoising diffusion probability model, stochastic processes, counterfactual inference, variational inference

CLC Number:

中图分类号：TP311.5

ZHANG Hongtao, ZHU Yongzhong, ZHANG Yuxuan, XIA Shiyuan. Causal Inference Algorithm Driven by Denoising Diffusion Probabilistic Model[J]. Computer and Modernization, 2025, 0(11): 119-126.

References

［1］ PEARL J. Causal inference in statistics: An overview［J］.Statistics Surveys， 2009（3）:96-146.
［2］ MORAFFAH R， MORAFFAH B， KARAMI M， et al. Causal adversarial network for learning conditional and interventional distributions［J］. arXiv preprint arXiv:2008.11
376， 2020.
［3］ PAWLOWSKI N， CASTRO D C， GLOCKER B. Deep structural causal models for tractable counterfactual inference［C］// Proceedings of the 34th International Conference on Neural Information Processing Systems. ACM， 2020: 857-869.
［4］ JUNG Y H， TIAN J， BAREINBOIM E. Learning causal effects via weighted empirical risk minimization［C］// Proceedings of the 34th International Conference on Neural Information Processing Systems. ACM， 2020:12697-12709.
［5］ ZE[C]EVICM， DHAMI D S， VELI[C]KOVIC P， et al. Relating graph neural networks to structural causal models［J］. arXiv preprint arXiv:2109.04173， 2021.
［6］ KHEMAKHEM I， MONTI R P， LEECH R， et al. Causal autoregressive flows［C］// International Conference on Artificial Intelligence and Statistics. PMLR， 2021: 3520-3528.
［7］ PARAFITA Á， VITRIÁ J. Causal inference with deep causal graphs［J］. arXiv preprint arXiv:2006.08380， 2020.
［8］ GARRIDO S， BORYSOV S S， RICH J， et al. Estimating causal effects with the neural autoregressive density estimator［J］. Journal of Causal Inference， 2021，9（1）:211-228.
［9］邓攀，刘俊廷，王晓，等. STCTN：一种基于时域偏倚校正与空域因果传递的时空因果表示学习方法［J］. 计算机学报， 2023，46（12）:2535-2550.
［10］ HO J， JAIN A， ABBEEL P. Denoising diffusion probabilistic models［C］// Proceedings of the 34th International Conference on Neural Information Processing Systems. ACM， 2020: 6840-6851.
［11］ ABSTREITER K， MITTAL S， BAUER S， et al. Diffusion-based representation learning［J］. arXiv preprint arXiv:2105.14257， 2021.
［12］ SANCHEZ P， LIU X， O'NEIL A Q， et al. Diffusion models for causal discovery via topological ordering［J］. arXiv preprint arXiv:2210.06201， 2022.
［13］ SANCHEZ P， TSAFTARIS S A. Diffusion causal models for counterfactual estimation［J］. arXiv preprint arXiv:2202.10166， 2022.
［14］ MUKHOPADHYAY S， GWILLIAM M， AGARWAL V， et al. Diffusion models beat GANs on image classification［J］. arXiv preprint arXiv:2307.08702， 2023.
［15］ CHAO P， BL[O]OBAUM P， KASIVISWANATHAN S P. Interventional and counterfactual inference with diffusion models［J］. arXiv preprint arXiv:2302.00860， 2023.
［16］ SHIMIZU T. Diffusion model in causal inference with unmeasured confounders［C］// 2023 IEEE Symposium Series on Computational Intelligence （SSCI）. IEEE， 2023:683-688.
［17］ KARIMI M A M， DITTADI A， BAUER S， et al. Diffusion-based causal representation learning［J］. Entropy， 2024， 26（7）: 556.
［18］蔡风景，蔡周策，李元. 基于DAG方法的经济变量因果关系研究［J］. 统计与信息论坛， 2008，23（9）:28-31.
［19］胡嘉伟. 结构因果模型中的do算子分析［J］. 自然辩证法通讯， 2024，46（6）:57-65.
［20］ SONG Y， SOHL-DICKSTEIN J， KINGMA D P， et al. Score-based generative modeling through stochastic differential equations［J］. arXiv preprint arXiv:2011.13456， 2020.
［21］ PEARL J， MACKENZIE D.The book of why: The new science of cause and effect［J］. Science， 2018，361（6405）:855.
［22］ BAREINBOIM E， CORREA J D， IBELING D， et al. On Pearl’s hierarchy and the foundations of causal inference［M］// Probabilistic and Causal Inference: The Works of Judea Pearl， ACM， 2022:507-556.
［23］ HANSEN N R， SOKOL A. Causal interpretation of stochastic differential equations［J］. Electronic Journal of Probability， 2014，19:1-24.
［24］ LUO C. Understanding diffusion models: A unified perspective［J］. arXiv preprint arXiv:2208.11970， 2022.
［25］ KARIMI A H， VON K[U]GELGEN J， SCH[O]LKOPF B， et al. Algorithmic recourse under imperfect causal knowledge: A probabilistic approach［J］.arXiv preprint arXiv:2006.06
831， 2022.
［26］ SANCHEZ-MARTIN P， RATEIKE M， VALERA I. VACA: Design of variational graph autoencoders for interventional and counterfactual queries［J］.arXiv preprint arXiv:2110.146
90， 2021.
［27］ LI H X， XIAO Y H， ZHENG C Y， et al. Balancing unobserved confounding with a few unbiased ratings in debiased recommendations［C］// Proceedings of the ACM Web Conference 2023. ACM， 2023:1305-1313.
［28］ LI H X， DAI Q Y， LI Y R， et al. Multiple robust learning for recommendation［C］// Proceedings of the AAAI Conference on Artificial Intelligence. AAAI， 2023，37（4）: 4417-4425.
［29］ LI H X， ZHENG C Y， WU P. StableDR: Stabilized doubly robust learning for recommendation on data missing not at random［J］. arXiv preprint arXiv:2205.04701， 2022.