Saved in:
Bibliographic Details
Main Authors: Muhammad, Omar, Shivkant, Pasupuleti Dhruv, Subramani, Deepak N.
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.07280
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915991745724416
author Muhammad, Omar
Shivkant, Pasupuleti Dhruv
Subramani, Deepak N.
author_facet Muhammad, Omar
Shivkant, Pasupuleti Dhruv
Subramani, Deepak N.
contents Leveraging deep learning for causal discovery in time series remains challenging because existing neural methods predominantly rely on component-wise architectures that fail to capture shared system dynamics or employ decoupled post-hoc graph extraction that risks overfitting to spurious correlations. We propose $\textbf{Mask2Cause}$, an end-to-end framework that recovers the underlying causal graph directly during the forecasting forward pass. Our approach introduces an Inverted Variable Embedding and an Adjacency-Constrained Masked Attention mechanism, trained with homoscedastic or heteroscedastic objectives to capture causal influences in both mean and variance. Empirical results on diverse benchmarks, from synthetic chaotic dynamics to realistic biological simulations, demonstrate state-of-the-art causal discovery with significantly reduced parameter complexity compared to standard baselines. We further show that inferred causal structures can be used to reduce parameter count of forecasting models by more than 70% on average while maintaining predictive accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2605_07280
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Mask2Cause: Causal Discovery via Adjacency Constrained Causal Attention
Muhammad, Omar
Shivkant, Pasupuleti Dhruv
Subramani, Deepak N.
Machine Learning
Artificial Intelligence
Leveraging deep learning for causal discovery in time series remains challenging because existing neural methods predominantly rely on component-wise architectures that fail to capture shared system dynamics or employ decoupled post-hoc graph extraction that risks overfitting to spurious correlations. We propose $\textbf{Mask2Cause}$, an end-to-end framework that recovers the underlying causal graph directly during the forecasting forward pass. Our approach introduces an Inverted Variable Embedding and an Adjacency-Constrained Masked Attention mechanism, trained with homoscedastic or heteroscedastic objectives to capture causal influences in both mean and variance. Empirical results on diverse benchmarks, from synthetic chaotic dynamics to realistic biological simulations, demonstrate state-of-the-art causal discovery with significantly reduced parameter complexity compared to standard baselines. We further show that inferred causal structures can be used to reduce parameter count of forecasting models by more than 70% on average while maintaining predictive accuracy.
title Mask2Cause: Causal Discovery via Adjacency Constrained Causal Attention
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2605.07280