[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86246-en":3,"doc-seo-86246-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},86246,687197207919,"Theodora","https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552",8,"Research & Report","CDFM: Towards a General-Purpose Causal Discovery Foundation Model","Causal discovery focuses on recovering causal structures from observational data, and prior methods often rely on dataset-specific workflows built around assumptions about the underlying causal mechanisms. As data grows in volume and heterogeneity, this assumption-specific, test-driven paradigm becomes fragmented and hard to scale. CDFM formulates a Causal Discovery Foundation Model as a unified zero-shot structural inference framework by analyzing identifiability boundaries, introducing causal priors, and using a variational decomposition that treats unknown mechanisms as latent variables. Pretraining on diverse synthetic structural causal models internalizes statistical asymmetries and improves performance over traditional approaches.","arXiv :2607 . 1 1508v 1 [ cs .LG] 13 Jul 2026  \nDMIR  \nLAB  \nCDFM: Towards a General-Purpose Causal Discovery  \nFoundation Model  \nJie Qiao 1 , Ruichu Cai 1∗ , Zijian Li 1 , Weilin Chen 1 , Pengfei Hua 1 , Boyan Xu 1 , Zhengming Chen2 , Zhifeng Hao2 , Peng Cui3  \n1 School of Computer Science, Guangdong University of Technology, Guangzhou, China  \n2 College of Mathematics and Computer, Shantou University, Shantou, China  \n3 Department of Computer Science and Technology, Tsinghua University, Beijing, China  \n{qiaojie.chn,cuiruichu,leizigin,chenweilin.chn,hua13662626240,hpakyim,[chenzhengming1103}@gmail.com](chenzhengming1103}@gmail.com) ,  \n[haozhifeng@stu.edu.cn](haozhifeng@stu.edu.cn) , [cuip@tsinghua.edu.cn](cuip@tsinghua.edu.cn)  \nAbstract  \nCausal 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.  \nCorrespondence: Ruichu Cai ([cairuichu@gmail.com](cairuichu@gmail.com))  \nCode: The code is available at [https://github.com/DMIRLAB-Group/CDFM](https://github.com/DMIRLAB-Group/CDFM).  \n1 Introduction  \nCausal discovery, the process of recovering underlying causal structures from observational data, is a fundamental pursuit across scientific disciplines, with broad applications in fields such as economics [9], biology [34], and social science [5] . Over the past decades, numerous causal discovery algorithms have been developed, demonstrating effectiveness across a wide variety of datasets [8, 20 , 33] .  \n(a) Traditional Causal Discovery Workflow  \n\n|  |\n| --- |\n| \u003Cbr>\u003Cbr>2 Data Preprocessing\u003Cbr>Missing Value Detection\u003Cbr>\u003Cbr>Select Imputation Method Mean KNN MICE ...\u003Cbr>\u003Cbr>Data Cleaning & Transformation\u003Cbr>\u003Cbr>3 Selecting Multiple Algorithms\u003Cbr> PC / FCI LiNGAM ANM LSNM\u003Cbr> |\n\n Paradigm Shift  \n\n|  |\n| --- |\n| \u003Cbr>\u003Cbr>\u003Cbr> Pretraining\u003Cbr>Diverse Mechanisms Causal Structures\u003Cbr>\u003Cbr> |\n\nFigure 1 . Paradigm shift from traditional causal discovery to CDFM. Instead of selecting algorithms and verifying assumptions for each dataset, CDFM performs lightweight preprocessing and directly infers causal structures using apretrained foundation model. The model is pretrained on diverse synthetic causal mechanisms and structures, enabling generalized causal reasoning across heterogeneous data distributions.  \nDespite these advancements, the field remains fund","cbCaiaswYFgObkig","https://ap.wps.com/l/cbCaiaswYFgObkig","pdf",1669507,3,1,27,"English","en",105,"# Introduction\n## Traditional causal discovery workflow\n## Paradigm shift toward CDFM","[{\"question\":\"What problem does CDFM target in causal discovery?\",\"answer\":\"CDFM targets the fragmented, test-driven causal discovery workflow that depends on fixed assumptions about causal mechanisms, which does not scale well to heterogeneous real-world data.\"},{\"question\":\"How does CDFM enable general-purpose causal structure inference?\",\"answer\":\"CDFM formulates causal discovery as a unified zero-shot structural inference framework, learning reusable inductive biases from diverse pretraining over synthetic causal mechanisms.\"},{\"question\":\"What theoretical and modeling components are used to handle unknown causal mechanisms?\",\"answer\":\"CDFM investigates causal identifiability boundaries, highlights the role of causal priors, and uses a principled variational framework that treats unknown causal mechanisms as latent variables and decomposes the marginal likelihood into tractable modules.\"}]",1784209779,68,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"cdfm-towards-a-general-purpose-causal-discovery-foundation-model","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/cdfm-towards-a-general-purpose-causal-discovery-foundation-model/86246/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does CDFM target in causal discovery?","Question",{"text":75,"@type":76},"CDFM targets the fragmented, test-driven causal discovery workflow that depends on fixed assumptions about causal mechanisms, which does not scale well to heterogeneous real-world data.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does CDFM enable general-purpose causal structure inference?",{"text":80,"@type":76},"CDFM formulates causal discovery as a unified zero-shot structural inference framework, learning reusable inductive biases from diverse pretraining over synthetic causal mechanisms.",{"name":82,"@type":73,"acceptedAnswer":83},"What theoretical and modeling components are used to handle unknown causal mechanisms?",{"text":84,"@type":76},"CDFM investigates causal identifiability boundaries, highlights the role of causal priors, and uses a principled variational framework that treats unknown causal mechanisms as latent variables and decomposes the marginal likelihood into tractable modules.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]