A Knowledge-Informed Pretrained Model for Causal Discovery

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Xu, Wenbo, He, Yue, Wang, Yunhai, Zhang, Xingxuan, Kuang, Kun, Chen, Yueguo, Cui, Peng
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908905388376064
author Xu, Wenbo
He, Yue
Wang, Yunhai
Zhang, Xingxuan
Kuang, Kun
Chen, Yueguo
Cui, Peng
author_facet Xu, Wenbo
He, Yue
Wang, Yunhai
Zhang, Xingxuan
Kuang, Kun
Chen, Yueguo
Cui, Peng
contents Causal discovery has been widely studied, yet many existing methods rely on strong assumptions or fall into two extremes: either depending on costly interventional signals or partial ground truth as strong priors, or adopting purely data driven paradigms with limited guidance, which hinders practical deployment. Motivated by real-world scenarios where only coarse domain knowledge is available, we propose a knowledge-informed pretrained model for causal discovery that integrates weak prior knowledge as a principled middle ground. Our model adopts a dual source encoder-decoder architecture to process observational data in a knowledge-informed way. We design a diverse pretraining dataset and a curriculum learning strategy that smoothly adapts the model to varying prior strengths across mechanisms, graph densities, and variable scales. Extensive experiments on in-distribution, out-of distribution, and real-world datasets demonstrate consistent improvements over existing baselines, with strong robustness and practical applicability.
format Preprint
id arxiv_https___arxiv_org_abs_2603_20842
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Knowledge-Informed Pretrained Model for Causal Discovery
Xu, Wenbo
He, Yue
Wang, Yunhai
Zhang, Xingxuan
Kuang, Kun
Chen, Yueguo
Cui, Peng
Machine Learning
Causal discovery has been widely studied, yet many existing methods rely on strong assumptions or fall into two extremes: either depending on costly interventional signals or partial ground truth as strong priors, or adopting purely data driven paradigms with limited guidance, which hinders practical deployment. Motivated by real-world scenarios where only coarse domain knowledge is available, we propose a knowledge-informed pretrained model for causal discovery that integrates weak prior knowledge as a principled middle ground. Our model adopts a dual source encoder-decoder architecture to process observational data in a knowledge-informed way. We design a diverse pretraining dataset and a curriculum learning strategy that smoothly adapts the model to varying prior strengths across mechanisms, graph densities, and variable scales. Extensive experiments on in-distribution, out-of distribution, and real-world datasets demonstrate consistent improvements over existing baselines, with strong robustness and practical applicability.
title A Knowledge-Informed Pretrained Model for Causal Discovery
topic Machine Learning
url https://arxiv.org/abs/2603.20842