notesum.ai

Published at December 5

Graph Disentangle Causal Model: Enhancing Causal Inference in Networked Observational Data

cs.LG

cs.IR

Released Date: December 5, 2024

Authors: Binbin Hu¹, Zhicheng An¹, Zhengwei Wu¹, Ke Tu¹, Ziqi Liu¹, Zhiqiang Zhang¹, Jun Zhou¹, Yufei Feng², Jiawei Chen³

Aff.: ¹Ant Group China, Hangzhou; ²Alibaba Group China, Hangzhou; ³Zhejiang University China, Hangzhou

Arxiv: http://arxiv.org/pdf/2412.03913v1

Refer to caption

	BlogCatalog						Flickr
$\kappa$	$\kappa$ =0.5		$\kappa$ =1		$\kappa$ =2		$\kappa$ =0.5		$\kappa$ =1		$\kappa$ =2
Method	$\sqrt{\epsilon_{PEHE}}$	$\epsilon_{ATE}$	$\sqrt{\epsilon_{PEHE}}$	$\epsilon_{ATE}$	$\sqrt{\epsilon_{PEHE}}$	$\epsilon_{ATE}$	$\sqrt{\epsilon_{PEHE}}$	$\epsilon_{ATE}$	$\sqrt{\epsilon_{PEHE}}$	$\epsilon_{ATE}$	$\sqrt{\epsilon_{PEHE}}$	$\epsilon_{ATE}$
BART	4.808	2.68	5.77	2.278	11.608	6.418	4.907	2.323	9.517	6.548	13.155	9.643
CF	7.456	1.261	7.805	1.763	19.271	4.05	8.104	1.359	14.636	3.545	26.702	4.324
TARNet	11.57	4.228	13.561	8.17	34.42	13.122	14.329	3.389	28.466	5.978	55.066	13.105
CFR-mmd	11.536	4.127	12.332	5.345	34.654	13.785	13.539	3.35	27.679	5.416	53.863	12.115
CFR-Wass	10.904	4.257	11.644	5.107	34.848	13.053	13.846	3.507	27.514	5.192	53.454	13.269
CEVAE	7.481	1.279	10.387	1.998	24.215	5.566	12.099	1.732	22.496	4.415	42.985	5.393
NetDeconf	4.532	0.979	4.597	0.984	9.532	2.130	4.286	0.805	5.789	1.359	9.817	2.700
GIAL	4.023	0.841	4.091	0.883	8.927	1.78	3.938	0.682	5.317	1.194	9.275	2.245
GNUM	4.122	0.932	4.367	0.973	9.327	2.082	4.102	0.801	5.345	1.267	9.756	2.675
IGL	3.987	0.815	4.012	0.845	8.746	1.628	4.012	0.703	5.437	1.259	9.312	2.321
DRCFR	3.772	0.698	3.857	1.070	6.077	1.405	4.019	0.703	5.367	1.304	7.915	2.691
GDC	3.012	0.236	3.499	0.591	5.941	0.876	3.952	0.281	5.351	0.491	8.287	1.241