Learning stable and predictive structures in kinetic systems: Benefits of a causal approach

Abstract

Learning kinetic systems from data is one of the core challenges in many fields. Identifying stable models is essential for the generalization capabilities of data-driven inference. We introduce a computationally efficient framework, called CausalKinetiX, that identifies structure from discrete time, noisy observations, generated from heterogeneous experiments. The algorithm assumes the existence of an underlying, invariant kinetic model, a key criterion for reproducible research. Results on both simulated and real-world examples suggest that learning the structure of kinetic systems benefits from a causal perspective. The identified variables and models allow for a concise description of the dynamics across multiple experimental settings and can be used for prediction in unseen experiments. We observe significant improvements compared to well established approaches focusing solely on predictive performance, especially for out-of-sample generalization.

References in corpus (15)

• Discovering governing equations from data: Sparse identification of nonlinear dynamical systems
• Approximate Bayesian computation scheme for parameter inference and model selection in dynamical systems
• Proceedings of the 29th International Conference on Machine Learning (ICML-12)
• Deep Hidden Physics Models: Deep Learning of Nonlinear Partial Differential Equations
• Structure Discovery in Nonparametric Regression through Compositional Kernel Search
• Iterative Random Forests to detect predictive and stable high-order interactions
• Veridical Data Science
• Model-free inference of direct network interactions from nonlinear collective dynamics
• On Causal and Anticausal Learning
• Learning stable and predictive structures in kinetic systems: Benefits of a causal approach
• From Ordinary Differential Equations to Structural Causal Models: the deterministic case
• From Deterministic ODEs to Dynamic Structural Causal Models
• Scalable Variational Inference for Dynamical Systems
• Learning Large Scale Ordinary Differential Equation Systems
• Reverse-engineering biological networks from large data sets

Cited by in corpus (11)

• Physics-informed learning of governing equations from scarce data
• Causality for Machine Learning
• Foundations of Structural Causal Models with Cycles and Latent Variables
• Anchor regression: heterogeneous data meets causality
• Causality in cognitive neuroscience: concepts, challenges, and distributional robustness
• Learning stable and predictive structures in kinetic systems: Benefits of a causal approach
• Distributional robustness of K-class estimators and the PULSE
• Switching Regression Models and Causal Inference in the Presence of Discrete Latent Variables
• Deconfounding and Causal Regularization for Stability and External Validity
• Orthogonal Structure Search for Efficient Causal Discovery from Observational Data
• Causality and Generalizability: Identifiability and Learning Methods