Amortized Causal Discovery: Learning to Infer Causal Graphs from Time-Series Data

Abstract

On time-series data, most causal discovery methods fit a new model whenever they encounter samples from a new underlying causal graph. However, these samples often share relevant information which is lost when following this approach. Specifically, different samples may share the dynamics which describe the effects of their causal relations. We propose Amortized Causal Discovery, a novel framework that leverages such shared dynamics to learn to infer causal relations from time-series data. This enables us to train a single, amortized model that infers causal relations across samples with different underlying causal graphs, and thus leverages the shared dynamics information. We demonstrate experimentally that this approach, implemented as a variational model, leads to significant improvements in causal discovery performance, and show how it can be extended to perform well under added noise and hidden confounding.

Accepted as a conference paper at CLeaR 2022

References in corpus (9)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• A Structured Self-attentive Sentence Embedding
• The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
• Interaction Networks for Learning about Objects, Relations and Physics
• Identifiability of parameters in latent structure models with many observed variables
• Causal Discovery in Physical Systems from Videos
• Neural Relational Inference with Fast Modular Meta-learning
• Discovering Nonlinear Relations with Minimum Predictive Information Regularization
• Amortized learning of neural causal representations

Cited by in corpus (14)

• Interpretable Models for Granger Causality Using Self-explaining Neural Networks
• Causal Inference for Time series Analysis: Problems, Methods and Evaluation
• Learning a Structural Causal Model for Intuition Reasoning in Conversation
• A Data-Driven Two-Phase Multi-Split Causal Ensemble Model for Time Series
• Transferable Time-Series Forecasting under Causal Conditional Shift
• Attention for Causal Relationship Discovery from Biological Neural Dynamics
• Causal Multi-Agent Reinforcement Learning: Review and Open Problems
• Learning interaction rules from multi-animal trajectories via augmented behavioral models
• EXplainable Neural-Symbolic Learning (X-NeSyL) methodology to fuse deep learning representations with expert knowledge graphs: the MonuMAI cultural heritage use case
• Self-supervised Reinforcement Learning with Independently Controllable Subgoals
• Dynamic Relational Inference in Multi-Agent Trajectories
• Inversion of Bayesian Networks
• Causal Discovery from Conditionally Stationary Time Series
• Convolutions for Spatial Interaction Modeling