Timezone: »
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 nonlinear 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 variables. 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.
Author Information
Yue Yu (Lehigh University)
Jie Chen (IBM Research)
Tian Gao (IBM Research)
Tian is currently a research staff member in IBM T. J. Watson Research Center. His research interests include machine learning, graphical models, causal discovery, reasoning, and applications.
Mo Yu (IBM T. J. Watson)
Related Events (a corresponding poster, oral, or spotlight)
-
2019 Poster: DAG-GNN: DAG Structure Learning with Graph Neural Networks »
Fri Jun 14th 01:30 -- 04:00 AM Room Pacific Ballroom
More from the Same Authors
-
2020 Poster: Invariant Rationalization »
Shiyu Chang · Yang Zhang · Mo Yu · Tommi Jaakkola -
2019 Poster: Generalized Linear Rule Models »
Dennis Wei · Sanjeeb Dash · Tian Gao · Oktay Gunluk -
2019 Oral: Generalized Linear Rule Models »
Dennis Wei · Sanjeeb Dash · Tian Gao · Oktay Gunluk -
2018 Poster: Parallel Bayesian Network Structure Learning »
Tian Gao · Dennis Wei -
2018 Oral: Parallel Bayesian Network Structure Learning »
Tian Gao · Dennis Wei -
2017 Poster: Local-to-Global Bayesian Network Structure Learning »
Tian Gao · Kshitij Fadnis · Murray Campbell -
2017 Talk: Local-to-Global Bayesian Network Structure Learning »
Tian Gao · Kshitij Fadnis · Murray Campbell