AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

CARL: Preserving Causal Structure in Representation Learning

Yulong Li, Xiwei Liu, Feilong Tang, Zhixiang Lu, Ming Hu, Yichen Li, Haochen Xue, Peixin Guo, Jionglong Su, Yutong Xie, Eran Segal, Imran Razzak

ICLR 2026 · Probabilistic and scientific modeling › Causal discovery · approach: Causal representation identifiability

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
add conditional independence loss L_CI (ablation)CARL · SyntheticCSI: 0.25 -> 1.00Improved
add conditional independence loss L_CI (ablation)CARL · SyntheticStructural accuracy: 0.40 -> 0.61Improved
add L_CI (ablation)CARL · SyntheticMAC: 0.83 -> 0.55Hurt
add Markov boundary preservation loss L_MBR (ablation)CARL · SyntheticMBRI: 0.46 -> 0.63Improved
add Markov boundary preservation loss L_MBR (ablation)CARL · SyntheticCSI: 0.75 -> 1.00Improved
add monotonic alignment loss L_MAC (ablation)CARL · SyntheticMAC: 0.32 -> 0.55Improved
add structure losses to alignment-only training (ablation)alignment-only · SyntheticCSI: 0.25 -> 1.00Improved
use independent predictor heads instead of shared (ablation)CARL · SyntheticCSI: 0.75 -> 1.0Improved
add cross-validation (ablation)CARL · SyntheticCSI: 0.25 -> 1.0Improved
use more negative samples than K = 32 (ablation)CARL · SyntheticCSI: 0.75 -> 1.0Improved
representation dimension d = 16 instead of full (ablation)CARL · Syntheticall metrics: within 0.05No clear effect
increase noise sigma from 0.1 to 0.5CARL · SyntheticMAC: 0.89 -> 0.42Hurt

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