Causal Discovery

15 papers across 5 sessions

Poster Session 1

1 paper

Wednesday, December 3, 2025 · 11:00 AM → 2:00 PM

Exhibit Hall C,D,E

Inexact Column Generation for Bayesian Network Structure Learning via Difference-of-Submodular Optimization

#700 · Yiran Yang, Rui Chen

Poster Session 2

2 papers

Wednesday, December 3, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

Characterization and Learning of Causal Graphs from Hard Interventions

#2514 · Zihan Zhou, Muhammad Qasim Elahi, Murat Kocaoglu

We introduce a graphical representation to characterize I-Markov Equivalence Class and propose a learning algorithm to integrate multiple datasets from hard interventions.

CausalDynamics: A large‐scale benchmark for structural discovery of dynamical causal models

#2517 · Benjamin Herdeanu, Juan Nathaniel, Carla Roesch, Jatan Buch, Gregor Ramien, Johannes Haux, Pierre Gentine

We present CausalDynamics, a benchmark for advancing the structural discovery of dynamical causal models consisting of true causal graphs derived from thousands of increasingly complex coupled differential equations.

Poster Session 3

1 paper

Thursday, December 4, 2025 · 11:00 AM → 2:00 PM

Exhibit Hall C,D,E

Score-informed Neural Operator for Enhancing Ordering-based Causal Discovery

#2606 · Jiyeon Kang, Songseong Kim, Chanhui Lee, Doyeong Hwang, Joanie Chung, Yunkyung Ko, Sumin Lee, Sungwoong Kim, Sungbin Lim

We enhance score matching based causal discovery algorithms through a specially designed neural operator trained via diffusion modeling in smooth function spaces.

Poster Session 5

9 papers

Friday, December 5, 2025 · 11:00 AM → 2:00 PM

Exhibit Hall C,D,E

When Additive Noise Meets Unobserved Mediators: Bivariate Denoising Diffusion for Causal Discovery

#2513 · Dominik Meier, Sujai Hiremath, PROMIT GHOSAL, Kyra Gan

We introduce a diffusion-based method for bivariate causal discovery that remains robust to unobserved mediators in additive noise models.

Differentiable Structure Learning and Causal Discovery for General Binary Data

#2514 · Chang Deng, Bryon Aragam

We propose a differentiable structure learning framework for general binary data that makes no parametric assumptions about the data‐generating process.

Less Greedy Equivalence Search

#2515 · Adiba Ejaz, Elias Bareinboim

We develop the Less Greedy Equivalence Search algorithm for learning causal structure from observational and interventional data with prior knowledge.

Local Learning for Covariate Selection in Nonparametric Causal Effect Estimation with Latent Variables

#2416 · Zheng Li, Xichen Guo, Feng Xie, Yan Zeng, Hao Zhang, Zhi Geng

We propose a novel local learning approach for covariate selection in nonparametric causal effect estimation, which accounts for the presence of latent variables.

Causal Mixture Models: Characterization and Discovery

#2415 · Sarah Mameche, Janis Kalofolias, Jilles Vreeken

Given a mixture of samples from unobserved subpopulations with distinct underlying causal mechanisms, we give results on identification and discovery of causal graph with latent mixing variables.

Practical Kernel Selection for Kernel-based Conditional Independence Test

#2411 · Wenjie Wang, Mingming Gong, Biwei Huang, James Bailey, Bo Han, Kun Zhang, Feng Liu

Pattern-Guided Adaptive Prior for Structure Learning

#2409 · Lyuzhou Chen, Yijia Sun, Yanze Gao, Xiangyu Wang, Derui Lyu, Taiyu Ban, Xin Wang, Xiren Zhou, Huanhuan Chen

Common prior knowledge can distort edge weights in DAG structure learning, creating characteristic graph patterns. Our framework detects these patterns as signals to adaptively adjust learning process, improving prior integration and DAG accuracy.

Differentiable Cyclic Causal Discovery Under Unmeasured Confounders

#2408 Spotlight · Muralikrishnna Guruswamy Sethuraman, Faramarz Fekri

We propose DCCD-CONF, a scalable, theoretically grounded framework for learning nonlinear cyclic causal graphs with unmeasured confounders using interventional data.

Differentiable Constraint-Based Causal Discovery

#2407 · Jincheng Zhou, Mengbo Wang, Anqi He, Yumeng Zhou, Hessam Olya, Murat Kocaoglu, Bruno Ribeiro

We propose a percolation-based differentiable d-separation framework that bridges constraint-based causal discovery methods with differentiable DAG structure learning.

Poster Session 6

2 papers

Friday, December 5, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

UnCLe: Towards Scalable Dynamic Causal Discovery in Non-linear Temporal Systems

#2507 · Tingzhu Bi, Yicheng Pan, Xinrui Jiang, Huize Sun, Meng Ma, Ping Wang

We propose UnCLe, a deep learning framework that uses temporal perturbations to uncover dynamic, time-varying causal graphs from observational time series.

Causal Discovery and Inference through Next-Token Prediction

#2510 · Eivinas Butkus, Nikolaus Kriegeskorte

Statistical prediction may be sufficient to drive the emergence of internal causal models and causal inference capacities in deep neural networks.