analiza aiqbench
This paper introduces ITSY, a novel continuous-optimization framework designed to perform causal discovery on irregular time-series data where traditional structural causal models fail due to missing observations. By jointly imputing missing values and learning graph structures using nearest-available history, the method demonstrates superior performance in recovering causal relationships under varying noise levels and missingness patterns. The work advances the field of machine learning by providing a robust solution for analyzing real-world temporal dynamics that are rarely perfectly sampled.
ITSY: Causal Discovery From Irregular Time-Series Data
arXiv:2609.23516v1 Announce Type: new Abstract: Structural causal models for time series recover contemporaneous and lagged effects, but most methods require complete observation windows and become misspecified when samples are missing. We introduce ITSY, the first continuous-optimization method for causal discovery from irregular time series under a linear model. ITSY reformulates the structural equation so that prediction uses the nearest available history rather than the possibly missing current slice, and jointly imputes missing values while learning both graphs. A weighted reconstruction objective corrects the noise transformation induced by this reformulation. Across synthetic regimes varying missingness, scale, graph density, and noise, and on a real world benchmark, ITSY consistently improves graph recovery over representative SCM-based baselines, demonstrating the effectiveness of the proposed method. The results establish a focused solution for irregular linear first-order dynamics and clarify the assumptions required for nonlinear or higher-order extensions.
Abstract:Structural causal models for time series recover contemporaneous and lagged effects, but most methods require complete observation windows and become misspecified when samples are missing. We introduce ITSY, the first continuous-optimization method for causal discovery from irregular time series under a linear model. ITSY reformulates the structural equation so that prediction uses the nearest available history rather than the possibly missing current slice, and jointly imputes missing values while learning both graphs. A weighted reconstruction objective corrects the noise transformation induced by this reformulation. Across synthetic regimes varying missingness, scale, graph density, and noise, and on a real world benchmark, ITSY consistently improves graph recovery over representative SCM-based baselines, demonstrating the effectiveness of the proposed method. The results establish a focused solution for irregular linear first-order dynamics and clarify the assumptions required for nonlinear or higher-order extensions.
Machine Learning (cs.LG)
arXiv:2609.23516 [cs.LG]
(or arXiv:2609.23516v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2609.23516 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 231 cuvinte
Subiecte conexe
Causal Discovery on Irregular Time Series
arXiv · 21 iul. 2026
Beyond Stationarity in Time Series: Discovering Causal Structures and Latent Regimes via Markov Blankets
arXiv · 7 sept. 2026
Learning Causal Structure of Time Series using Best Order Score Search
arXiv · 26 sept. 2026
Causal Discovery with Inverted Self-attention for Multivariate Time Series
arXiv · 31 iul. 2026