DAG-GNN: DAG Structure Learning with Graph Neural Networks

Learning a faithful directed acyclic graph (DAG) from samples of a joint distribution is a challenging combinatorial problem, owing to the intractable search space superexponential in the number of graph nodes. A recent breakthrough formulates the problem as a continuous optimization with a structural constraint that ensures acyclicity (Zheng et al., 2018). The authors apply the approach to the linear structural equation model (SEM) and the least-squares loss function that are statistically well justified but nevertheless limited. Motivated by the widespread success of deep learning that is capable of capturing complex nonlinear mappings, in this work we propose a deep generative model and apply a variant of the structural constraint to learn the DAG. At the heart of the generative model is a variational autoencoder parameterized by a novel graph neural network architecture, which we coin DAG-GNN. In addition to the richer capacity, an advantage of the proposed model is that it naturally handles discrete variables as well as vector-valued ones. We demonstrate that on synthetic data sets, the proposed method learns more accurate graphs for nonlinearly generated samples; and on benchmark data sets with discrete variables, the learned graphs are reasonably close to the global optima. The code is available at \url{this https URL}.

• Publication year

  2019

• Venue

  International Conference on Machine Learning

• Publication date

  2019-04-22

• Fields of study

  Mathematics, Computer Science

• Identifiers
  arXiv 1904.10098
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Nonlinear Programming

  2021influential reference
• DAGs with NO TEARS: Continuous Optimization for Structure Learning

  2018influential reference
• Parallel Bayesian Network Structure Learning

  2018cited by this paper
• FastGCN: Fast Learning with Graph Convolutional Networks via Importance Sampling

  2018influential reference
• SAM: Structural Agnostic Model, Causal Discovery and Penalized Adversarial Learning

  2018cited by this paper
• Graph Attention Networks

  2017cited by this paper
• Inductive Representation Learning on Large Graphs

  2017influential reference
• Automatic differentiation in PyTorch

  2017cited by this paper
• Neural Message Passing for Quantum Chemistry

  2017influential reference
• Learning Bayesian networks with ancestral constraints

  2016cited by this paper
• A million variables and more: the Fast Greedy Equivalence Search algorithm for learning high-dimensional graphical causal models, with an application to functional magnetic resonance images

  2016cited by this paper
• Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering

  2016influential reference
• Joint Causal Inference from Multiple Contexts

  2016cited by this paper
• Semi-Supervised Classification with Graph Convolutional Networks

  2016cited by this paper
• Representing Text for Joint Embedding of Text and Knowledge Bases

  2015cited by this paper
• Learning Bayesian Networks with Thousands of Variables

  2015cited by this paper
• Polyhedral aspects of score equivalence in Bayesian network structure learning

  2015influential reference
• Gated Graph Sequence Neural Networks

  2015influential reference
• Structure Learning in Bayesian Networks of a Moderate Size by Efficient Sampling

  2015cited by this paper
• Advances in Learning Bayesian Networks of Bounded Treewidth

  2014cited by this paper
• Adam: A Method for Stochastic Optimization

  2014cited by this paper
• Annealed Importance Sampling for Structure Learning in Bayesian Networks

  2013cited by this paper
• Auto-Encoding Variational Bayes

  2013influential reference
• Spectral Networks and Locally Connected Networks on Graphs

  2013influential reference
• Learning Optimal Bayesian Networks: A Shortest Path Perspective

  2013cited by this paper
• Structured Learning of Gaussian Graphical Models

  2012cited by this paper
• Bayesian Structure Learning in Sparse Gaussian Graphical Models

  2012cited by this paper
• Partial Order MCMC for Structure Discovery in Bayesian Networks

  2011cited by this paper
• Bayesian network learning with cutting planes

  2011influential reference
• Learning Bayesian networks by hill climbing: efficient methods based on progressive restriction of the neighborhood

  2010cited by this paper
• Learning Bayesian Network Structure using LP Relaxations

  2010influential reference
• On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias

  2008cited by this paper
• Improving the structure MCMC sampler for Bayesian networks by introducing a new edge reversal move

  2008cited by this paper
• Bayesian structure learning using dynamic programming and MCMC

  2007cited by this paper
• A Simple Approach for Finding the Globally Optimal Bayesian Network Structure

  2006cited by this paper
• The max-min hill-climbing Bayesian network structure learning algorithm

  2006cited by this paper
• Finding optimal Bayesian networks by dynamic programming

  2005cited by this paper
• Causal Protein-Signaling Networks Derived from Multiparameter Single-Cell Data

  2005cited by this paper
• Exact Bayesian Structure Discovery in Bayesian Networks

  2004cited by this paper
• Being Bayesian About Network Structure. A Bayesian Approach to Structure Discovery in Bayesian Networks

  2004cited by this paper
• Finding Optimal Models for Small Gene Networks

  2003cited by this paper
• Time and sample efficient discovery of Markov blankets and direct causal relations

  2003cited by this paper
• Causation, Prediction, and Search

  2003cited by this paper
• Optimal Structure Identification With Greedy Search

  2002influential reference
• Large-Sample Learning of Bayesian Networks is NP-Hard

  2002cited by this paper
• Constructing Bayesian Network Models of Gene Expression Networks from Microarray Data

  2000cited by this paper
• Computation, causation, and discovery

  2000cited by this paper
• Data Analysis with Bayesian Networks: A Bootstrap Approach

  1999cited by this paper
• Bayesian Graphical Models for Discrete Data

  1995cited by this paper
• Causal Inference in the Presence of Latent Variables and Selection Bias

  1995cited by this paper
• Learning Bayesian networks: The combination of knowledge and statistical data

  1994cited by this paper
• Probabilistic reasoning in intelligent systems: Networks of plausible inference

  1991influential reference
• Approximating discrete probability distributions with dependence trees

  1968cited by this paper
• Causality : Models , Reasoning , and Inference

  year unknowncited by this paper

• Dynamic Expert-Guided Model Averaging for Causal Discovery

  2026cites this paper
• Causal-Guided Strength Differential Independence Sample Weighting for Out-of-Distribution Generalization

  2026cites this paper
• MSDG: A multiscale dynamic graph Neural network for inferring dynamic granger causality in multivariate time series

  2026cites this paper
• Efficient Differentiable Causal Discovery via Reliable Super-Structure Learning

  2026cites this paper
• Causal Structure-Enhanced Branch Neural Networks for Interpretable and Robust Regression

  2026cites this paper
• Reinforced Direct Causality Learning for Enhanced Fault Path Diagnosis

  2026cites this paper
• Multi-channel causal variational autoencoder for multimodal biomedical causal disentanglement.

  2026cites this paper
• Causality-Informed Models for Multidrug-Resistant Organism Prediction: Enabling Effective ICU Antibiotic Stewardship.

  2026cites this paper
• Temporal Causal Graph-based Attention Gated Recurrent Unit for Interpretable Fault Diagnosis in Nuclear Power Plants

  2026cites this paper
• Root cause analysis and prediction of quality issues in the complex aviation products assembly process using object oriented Bayesian networks

  2026cites this paper
• CVAE-based Causal Representation Learning from Retinal Fundus Images for Age Related Macular Degeneration(AMD) Prediction

  2026cites this paper
• Semi-Supervised and Disentangled Causal Discovery for Analyzing Fault Propagation in Microservices

  2026influential citation
• CRFD: Causal and Interpretable Fault Diagnosis Using Counterfactual Reasoning for Microservices on Multi-source Observability Data

  2026cites this paper
• DKC: Data-driven and knowledge-guided causal discovery with application to healthcare data

  2026cites this paper
• LungCRCT: Causal Representation based Lung CT Processing for Lung Cancer Treatment

  2026cites this paper
• A Causal Perspective for Enhancing Jailbreak Attack and Defense

  2026cites this paper
• CAMEL-Dx: A Causal Meta-Learning Framework for Rare Disease Diagnosis with Limited Samples

  2026cites this paper
• Causal-enhanced Machine Learning Framework for Long-term Wind Power Probabilistic Forecasting

  2026cites this paper
• Federated Learning of Nonlinear Temporal Dynamics with Graph Attention-based Cross-Client Interpretability

  2026cites this paper
• SC3D: Dynamic and Differentiable Causal Discovery for Temporal and Instantaneous Graphs

  2026cites this paper
• Ordering-based Causal Discovery via Generalized Score Matching

  2026cites this paper
• Heterogeneous-Effect Causal Graph Diffusion Network for sustainable regional economic upgrading

  2026cites this paper
• Efficient Causal Structure Learning via Modular Subgraph Integration

  2026cites this paper
• Self-supervised session-based recommendation model based on global causal structure learning

  2026cites this paper
• Recovering Whole-Brain Causal Connectivity under Indirect Observation with Applications to Human EEG and fMRI

  2026cites this paper
• Test-Time Learning of Causal Structure from Interventional Data

  2026cites this paper
• Coarse-to-Fine Learning of Dynamic Causal Structures

  2026cites this paper
• DFT AI: Machine Learning-Guided Test Point Insertion for Pre-Silicon Debug and Post-Silicon Diagnosis in Complex VLSI Systems

  2026influential citation
• A causal generative model-based optimal scheduling method for blast furnace gas system considering unknown scenarios

  2026cites this paper
• TabSynergy: Leveraging large language models for high-quality feature generation and augmentation in small tabular datasets

  2025cites this paper
• Spatially varying cell-specific gene regulation network inference

  2025cites this paper
• An improved greedy equivalent search method based on relative entropy

  2025cites this paper
• Causal Graph-Based Root Cause Analysis for ICS Sensor Streams

  2025cites this paper
• Score-based Greedy Search for Structure Identification of Partially Observed Linear Causal Models

  2025cites this paper
• CausalMamba: Scalable Conditional State Space Models for Neural Causal Inference

  2025cites this paper
• AKCD: An Attention-Driven Framework with Kernel Conditional Independence Testing for Time Series Causal Discovery

  2025cites this paper
• Improving constraint-based discovery with robust propagation and reliable LLM priors

  2025cites this paper
• Causality-Induced Positional Encoding for Transformer-Based Representation Learning of Non-Sequential Features

  2025influential citation
• Robust and Lightweight Root Cause Analysis for Microservice Metrics

  2025cites this paper
• Causal Discovery via Quantile Partial Effect

  2025cites this paper
• Effects of Distributional Biases on Gradient-Based Causal Discovery in the Bivariate Categorical Case

  2025cites this paper
• Generative Causality-Driven Network for Graph Multi-Task Learning

  2025cites this paper
• DAG DECORation: Continuous Optimization for Structure Learning under Hidden Confounding

  2025cites this paper
• Shylock: Causal Discovery in Multivariate Time Series based on Hybrid Constraints

  2025cites this paper
• Bayesian Inference for High-dimensional Time Series with a Stationary Directed Acyclic Graphical Structure

  2025cites this paper
• Learning Causal Graphs at Scale: A Foundation Model Approach

  2025cites this paper
• Graph Concept Bottleneck Models

  2025cites this paper
• A deep spatio-temporal architecture for dynamic ECN analysis with Granger causality based causal discovery

  2025cites this paper
• Traffic prediction using an active causality recurrent graph convolutional network

  2025cites this paper
• PersonaX: Multimodal Datasets with LLM-Inferred Behavior Traits

  2025cites this paper
• HiSpTRSV: Exploring Tile-Level Parallelism for SpTRSV Acceleration on FPGAs

  2025cites this paper
• Towards Xai for the Facilitation of Human-Machine Learning of Graph Structure

  2025cites this paper
• Novel graph neural network and GNN-C-Transformer model construction for direction of arrival estimation

  2025cites this paper
• Differentiable Structure Learning and Causal Discovery for General Binary Data

  2025cites this paper
• Bridging insight gaps in topic dependency discovery with a knowledge-inspired topic model

  2025cites this paper
• GCVAMD: A Modified CausalVAE Model for Causal Age-related Macular Degeneration Risk Factor Detection and Prediction

  2025cites this paper
• UnG-MoCha: Neural Motif Counting in Uncertain Graphs

  2025cites this paper
• CALM: A Causal Analysis Language Model for Tabular Data in Complex Systems with Local Scores, Conditional Independence Tests, and Relation Attributes

  2025cites this paper
• Competition is the key: A Game Theoretic Causal Discovery Approach

  2025cites this paper
• Causal similarity learning with multi-level predictive relation aggregation for grouped root cause diagnosis of industrial faults

  2025cites this paper
• Graph Autoencoder and StrNN based Causal Analysis of Mortality in Heart Failure Patients

  2025influential citation
• Flippable Siamese Differential Neural Network for Differential Graph Inference

  2025cites this paper
• Causal Graph Generation and Validation for Cognitive 6G Networks

  2025cites this paper
• TV-CCANM: a transformer variational inference in confounding cascade additive noise model for causal effect estimation

  2025cites this paper
• Topological-Temporal Convolution Transformer Hawkes Process for Causal Structural Learning in Telecom Networks

  2025influential citation
• Active differentiable structure learning for clinical causal discovery

  2025cites this paper
• Directed Graph Grammars for Sequence-based Learning

  2025cites this paper
• Inferring Directed Acyclic Graphs from Event Sequences via Learning Gromov-Wasserstein-Regularized Hawkes Processes

  2025cites this paper
• Hand Drawing Image based Causal Representation Learning for Robust Parkinson’s Disease Feature Extraction and Detection

  2025cites this paper
• CausalConf: Datasize-Aware Configuration Auto-Tuning for Recurring Big Data Processing Jobs via Adaptive Causal Structure Learning

  2025cites this paper
• Denoising dual sparse graph attention model for session-based recommendation

  2025cites this paper
• Integrating machine learning with triboelectric nanogenerators: Optimizing electrode materials and doping strategies for intelligent energy harvesting

  2025cites this paper
• Nonlinear Causal Discovery through a Sequential Edge Orientation Approach

  2025cites this paper
• Pseudo-intervention based local-to-global causal structure learning

  2025cites this paper
• Analytic DAG Constraints for Differentiable DAG Learning

  2025cites this paper
• Integrating causal representations with domain adaptation for fault diagnosis

  2025cites this paper
• An information-theoretic approach for heterogeneous differentiable causal discovery

  2025cites this paper
• CAUSAL3D: A Comprehensive Benchmark for Causal Learning from Visual Data

  2025cites this paper
• Causal Discovery from Shifted Multiple Environments

  2025cites this paper
• Causal discovery based on hierarchical reinforcement learning

  2025cites this paper
• Inferring gene regulatory networks by hypergraph generative model

  2025cites this paper
• A conceptual synthesis of causal assumptions for causal discovery and inference

  2025cites this paper
• Inexact Column Generation for Bayesian Network Structure Learning via Difference-of-Submodular Optimization

  2025cites this paper
• Adaptive hierarchical knowledge distillation from GNNs to MLPs

  2025cites this paper
• Bayesian network structure learning by opposition-based learning

  2025cites this paper
• Generating Hypotheses of Dynamic Causal Graphs in Neuroscience: Leveraging Generative Factor Models of Observed Time Series

  2025cites this paper
• From Images to Insights: Explainable Biodiversity Monitoring with Plain Language Habitat Explanations

  2025cites this paper
• Time-Reversal Enhanced Dynamic Causality Distribution Learning and Its Application in Identifying Dynamic ECNs in MCI Patients

  2025cites this paper
• ABC: Adaptive BayesNet Structure Learning for Computational Scalable Multi-task Image Compression

  2025cites this paper
• Causality-Aided Evaluation and Explanation of Large Language Model-Based Code Generation

  2025cites this paper
• GNNMG-PPO for Deployment Optimization of Cell-Free Access Points

  2025cites this paper
• The robustness of differentiable Causal Discovery in misspecified Scenarios

  2025influential citation
• From Observations to Causations: A GNN-based Probabilistic Prediction Framework for Causal Discovery

  2025influential citation
• AmocRCA: At Most One Change Segmentation and Relative Correlation Ranking for Root Cause Analysis

  2025cites this paper
• Optimized Consensus with DAGWise: A GNN-Enhanced Approach for Scalable and Fault-Tolerant DAG-Based BFT

  2025cites this paper
• Score-informed Neural Operator for Enhancing Ordering-based Causal Discovery

  2025cites this paper
• DHS-PE: A More Suitable Positional Encoding for Directed Hypergraphs

  2025cites this paper
• MOCHA: Discovering Multi-Order Dynamic Causality in Temporal Point Processes

  2025cites this paper
• Uncertain Priors for Graphical Causal Models: A Multi-Objective Optimization Perspective

  2025cites this paper
• InfoSEM: A Deep Generative Model with Informative Priors for Gene Regulatory Network Inference

  2025cites this paper