DAGs with No Fears: A Closer Look at Continuous Optimization for Learning Bayesian Networks

Abstract

This paper re-examines a continuous optimization framework dubbed NOTEARS for learning Bayesian networks. We first generalize existing algebraic characterizations of acyclicity to a class of matrix polynomials. Next, focusing on a one-parameter-per-edge setting, it is shown that the Karush-Kuhn-Tucker (KKT) optimality conditions for the NOTEARS formulation cannot be satisfied except in a trivial case, which explains a behavior of the associated algorithm. We then derive the KKT conditions for an equivalent reformulation, show that they are indeed necessary, and relate them to explicit constraints that certain edges be absent from the graph. If the score function is convex, these KKT conditions are also sufficient for local minimality despite the non-convexity of the constraint. Informed by the KKT conditions, a local search post-processing algorithm is proposed and shown to substantially and universally improve the structural Hamming distance of all tested algorithms, typically by a factor of 2 or more. Some combinations with local search are both more accurate and more efficient than the original NOTEARS.

arXiv comment40 pages, 8 figures, to appear at the 34th Conference on Neural Information Processing Systems (NeurIPS 2020)

Cited by in corpus (11)

• Causality-based CTR Prediction using Graph Neural Networks
• Scalable Causal Structure Learning: Scoping Review of Traditional and Deep Learning Algorithms and New Opportunities in Biomedicine
• Towards Human-like Perception: Learning Structural Causal Model in Heterogeneous Graph
• Sparse Bayesian Estimation of Parameters in Linear-Gaussian State-Space Models
• Beware of the Simulated DAG! Causal Discovery Benchmarks May Be Easy To Game
• Towards Federated Bayesian Network Structure Learning with Continuous Optimization
• Discovering Dynamic Causal Space for DAG Structure Learning
• On the Convergence of Continuous Constrained Optimization for Structure Learning
• Fast Causal Discovery by Approximate Kernel-based Generalized Score Functions with Linear Computational Complexity
• Inference of Causal Effects when Control Variables are Unknown
• Identifiability of AMP chain graph models