Learning Temporally Causal Latent Processes from General Temporal Data

2021/10/11 by Weiran Yao, Yuewen Sun, Yao, Weiran +8 · 12 citations
Computer Science · Mathematics · #Artificial Intelligence (cs.AI) #Bayesian Modeling and Causal Inference #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Time Series Analysis and Forecasting #cs.AI #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.2110.05428

ICLR 2022: https://openreview.net/forum?id=RDlLMjLJXdq

openalex publication_date 2021/10/11 · arxiv created 2022/02/08 · arxiv updated 2022/02/10 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Our goal is to recover time-delayed latent causal variables and identify their relations from measured temporal data. Estimating causally-related latent variables from observations is particularly challenging as the latent variables are not uniquely recoverable in the most general case. In this work, we consider both a nonparametric, nonstationary setting and a parametric setting for the latent processes and propose two provable conditions under which temporally causal latent processes can be identified from their nonlinear mixtures. We propose LEAP, a theoretically-grounded framework that extends Variational AutoEncoders (VAEs) by enforcing our conditions through proper constraints in causal process prior. Experimental results on various datasets demonstrate that temporally causal latent processes are reliably identified from observed variables under different dependency structures and that our approach considerably outperforms baselines that do not properly leverage history or nonstationarity information. This demonstrates that using temporal information to learn latent processes from their invertible nonlinear mixtures in an unsupervised manner, for which we believe our work is one of the first, seems promising even without sparsity or minimality assumptions.

Citations

• An Algorithm for Fast Recovery of Sparse Causal Graphs
• A Meta-Transfer Objective for Learning to Disentangle Causal Mechanisms
• Unsupervised Learning of Object Keypoints for Perception and Control
• Learning Neural Causal Models from Unknown Interventions
• Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Coding
• Generalized Independent Noise Condition for Estimating Latent Variable Causal Graphs

Cited by

• LacaDM: A Latent Causal Diffusion Model for Multiobjective Reinforcement Learning
• FlowNet: Modeling Dynamic Spatio-Temporal Systems via Flow Propagation
• Debiasing Reward Models by Representation Learning with Guarantees
• Thought Communication in Multiagent Collaboration
• Towards Identifiability of Hierarchical Temporal Causal Representation Learning
• Online Time Series Forecasting with Theoretical Guarantees
• CausalVerse: Benchmarking Causal Representation Learning with Configurable High-Fidelity Simulations
• LLM Interpretability with Identifiable Temporal-Instantaneous Representation
• Incorporating Interventional Independence Improves Robustness against Interventional Distribution Shift
• Causal Spatio-Temporal Prediction: An Effective and Efficient Multi-Modal Approach
• Modeling Unseen Environments with Language-guided Composable Causal Components in Reinforcement Learning
• Causal View of Time Series Imputation: Some Identification Results on Missing Mechanism

Related

• Learning Latent Causal Dynamics
• Temporally Disentangled Representation Learning under Unknown Nonstationarity
• Temporally Disentangled Representation Learning
• CaRiNG: Learning Temporal Causal Representation under Non-Invertible Generation Process
• On the Identification of Temporally Causal Representation with Instantaneous Dependence