ProDAG: Projected Variational Inference for Directed Acyclic Graphs
Not on the map. Why?
Abstract (arXiv)
Directed acyclic graph (DAG) learning is a central task in structure discovery and causal inference. Although the field has witnessed remarkable advances over the past few years, it remains statistically and computationally challenging to learn a single (point estimate) DAG from data, let alone provide uncertainty quantification. We address the difficult task of quantifying graph uncertainty by developing a Bayesian variational inference framework based on novel, provably valid distributions that have support directly on the space of sparse DAGs. These distributions, which we use to define our prior and variational posterior, are induced by a projection operation that maps an arbitrary continuous distribution onto the space of sparse weighted acyclic adjacency matrices. While this projection is combinatorial, it can be solved efficiently using recent continuous reformulations of acyclicity constraints. We empirically demonstrate that our method, ProDAG, can outperform state-of-the-art alternatives in both accuracy and uncertainty quantification.
Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 32 of its 46 references that are in the map. First posted on arXiv 24 May 2024.
Builds on
32 of 46 references.
References · 32
- 2014Adam: A Method for Stochastic Optimizationcited by 37,240
- 2016Variational Inference: A Review for Statisticianscited by 1,472
- 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 752
- 2020Differentiable Causal Discovery from Interventional Datacited by 218
- 2021D'ya like DAGs? A Survey on Structure Learning and Causal Discoverycited by 211
- 2019Learning Sparse Nonparametric DAGscited by 207
- 2020On the Role of Sparsity and DAG Constraints for Learning Linear DAGscited by 191
- 2022DAGMA: Learning DAGs via M-matrices and a Log-Determinant Acyclicity Characterizationcited by 140
- 2021DiBS: Differentiable Bayesian Structure Learningcited by 106
- 2022Deep End-to-end Causal Inferencecited by 95
- 2021A survey of Bayesian Network structure learningcited by 87
- 2021BCD Nets: Scalable Variational Approaches for Bayesian Causal Discoverycited by 68
20 more
- 2018Fashionable Modelling with Fluxcited by 53
- 2021DAGs with No Curl: An Efficient DAG Structure Learning Approachcited by 53
- 2023Joint Bayesian Inference of Graphical Structure and Parameters with a Single Generative Flow Networkcited by 50
- 2022Differentiable DAG Samplingcited by 47
- 2022Large-Scale Differentiable Causal Discovery of Factor Graphscited by 47
- 2021Variational Causal Networks: Approximate Bayesian Inference over Causal Structurescited by 43
- 2023Fast Scalable and Accurate Discovery of DAGs Using the Best Order Score Search and Grow-Shrink Treescited by 38
- 2020Towards Scalable Bayesian Learning of Causal DAGscited by 33
- 2023BayesDAG: Gradient-Based Posterior Inference for Causal Discoverycited by 33
- 2022Differentiable Causal Discovery Under Latent Interventionscited by 21
- 2019Integer Programming for Learning Directed Acyclic Graphs from Continuous Datacited by 20
- 2023Optimizing NOTEARS Objectives via Topological Swapscited by 17
- 2022Tractable Uncertainty for Structure Learningcited by 15
- 2023Differentiable Multi-Target Causal Bayesian Experimental Designcited by 15
- 2024Effective Bayesian Causal Inference via Structural Marginalisation and Autoregressive Orderscited by 6
- 2024Permutation-based Inference for Variational Learning of Directed Acyclic Graphscited by 4
- 2023Contextual Directed Acyclic Graphscited by 3
- 2021Learning Large DAGs by Combining Continuous Optimization and Feedback Arc Set Heuristicscited by 2
- 2020Bayesian Inference with the l1-ball Prior: Solving Combinatorial Problems with Exact Zeroscited by 1
- 2021Bayesian Inference using the Proximal Mapping: Uncertainty Quantification under Varying Dimensionalitycited by 1
Built on by
Citing papers per year: 2026: 4.
4 papers.