AI Paper Atlas

Research map / Probabilistic and scientific modeling / Treatment effects

Causal Effect Estimation Under Networked Interference Without Networked Unconfoundedness Assumption

Weilin Chen, Ruichu Cai, Jie Qiao, Yuguang Yan, José Miguel Hernández-Lobato

TPAMI 2026 · Probabilistic and scientific modeling › Treatment effects · approach: Latent-variable causal inference · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
recover role-specific latent confounders (SPNet+z has no recovery guarantee) instead of SPNet+zCaLaNet · BC(homo) within sampleATE error: 0.0570 -> 0.0288Improved
CaLaNet versus strongest baseline TNet on ATECaLaNet · BC(hete) within sampleATE error: 0.0429 -> 0.0401No clear effect
CaLaNet versus RRNet on ATECaLaNet · Flickr(homo) within sampleATE error: 0.0199 -> 0.0151Improved
CaLaNet versus TNet on ITECaLaNet · BC(hete) within sampleITE error: 0.1458 -> 0.1481Hurt
use very small alpha (0.01) for the IPM balancing weight (ablation)CaLaNet · BC(homo)/Flickr(homo) individual errorindividual effect errorHurt
use large beta (10.0) for the outcome loss weight (ablation)CaLaNet · BC(homo)/Flickr(homo) individual errorindividual effect errorHurt
vary alpha and beta within moderate range over 0.01, 0.1, 1.0, 10.0 (ablation)CaLaNet · BC(homo)/Flickr(homo)individual effect errorNo clear effect
add recovery of three latent confounder types (u^i, u^c, u^n) via identifiable representation learningCaLaNet · SyntheticMCC: 0.9729 vs 0.1742 (U^i vs off-diagonal)Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy