On the Sparse DAG Structure Learning Based on Adaptive Lasso
Not on the map. Why?
Abstract (arXiv)
Learning the underlying Bayesian Networks (BNs), represented by directed acyclic graphs (DAGs), of the concerned events from purely-observational data is a crucial part of evidential reasoning. This task remains challenging due to the large and discrete search space. A recent flurry of developments followed NOTEARS[1] recast this combinatorial problem into a continuous optimization problem by leveraging an algebraic equality characterization of acyclicity. However, the continuous optimization methods suffer from obtaining non-spare graphs after the numerical optimization, which leads to the inflexibility to rule out the potentially cycle-inducing edges or false discovery edges with small values. To address this issue, in this paper, we develop a completely data-driven DAG structure learning method without a predefined value to post-threshold small values. We name our method NOTEARS with adaptive Lasso (NOTEARS-AL), which is achieved by applying the adaptive penalty method to ensure the sparsity of the estimated DAG. Moreover, we show that NOTEARS-AL also inherits the oracle properties under some specific conditions. Extensive experiments on both synthetic and a real-world dataset demonstrate that our method consistently outperforms NOTEARS.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 27 of its 58 references that are in the map. First posted on arXiv 7 Sep 2022.
Builds on
27 of 58 references.
References · 27
- 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 752
- 2019DAG-GNN: DAG Structure Learning with Graph Neural Networkscited by 331
- 2019Causality for Machine Learningcited by 243
- 2019Gradient-Based Neural DAG Learningcited by 227
- 2020Differentiable Causal Discovery from Interventional Datacited by 218
- 2019Learning Sparse Nonparametric DAGscited by 207
- 2020DYNOTEARS: Structure Learning from Time-Series Datacited by 200
- 2020On the Role of Sparsity and DAG Constraints for Learning Linear DAGscited by 191
- 2013Learning Bayesian Networks: The Combination of Knowledge and Statistical Datacited by 187
- 2019Causal Discovery with Reinforcement Learningcited by 145
- 2022Deep End-to-end Causal Inferencecited by 95
- 2019Masked Gradient-Based Causal Structure Learningcited by 89
15 more
- 2020DAGs with No Fears: A Closer Look at Continuous Optimization for Learning Bayesian Networkscited by 66
- 2013Learning Gaussian Networkscited by 62
- 2020A Tutorial on Learning With Bayesian Networkscited by 57
- 2020Differentiable Causal Discovery Under Unmeasured Confoundingcited by 56
- 2018Structural Agnostic Modeling: Adversarial Learning of Causal Graphscited by 48
- 2020On the Convergence of Continuous Constrained Optimization for Structure Learningcited by 38
- 2021Towards Federated Bayesian Network Structure Learning with Continuous Optimizationcited by 31
- 2020Causality Learning: A New Perspective for Interpretable Machine Learningcited by 28
- 2022Truncated Matrix Power Iteration for Differentiable DAG Learningcited by 23
- 2022MissDAG: Causal Discovery in the Presence of Missing Data with Continuous Additive Noise Modelscited by 22
- 2022Reliable Causal Discovery with Improved Exact Search and Weaker Assumptionscited by 19
- 2021FedDAG: Federated DAG Structure Learningcited by 17
- 2020Targeted VAE: Variational and Targeted Learning for Causal Inferencecited by 7
- 2014Proximal Quasi-Newton for Computationally Intensive L1-regularized M-estimatorscited by 2
- 2013The restricted consistency property of leave-nv-out cross-validation for high-dimensional variable selectioncited by 1
Built on by
Citing papers per year: 2023: 2; 2024: 0; 2025: 1.
3 papers.