Nonnegative DAG Learning via Concomitant Estimation
Abstract
We study the problem of learning directed acyclic graphs (DAGs) with nonnegative edge weights from observational data. We propose the Nonnegative and Concomitant (NoCo) DAG estimator, which jointly recovers the weighted graph structure and the exogenous noise variances in the linear structural equation model for the observations. Different from prior art, this noise-adaptive formulation blends a smoothed concomitant lasso criterion with a simpler log-determinant acyclicity constraint that exploits nonnegativity and yields a more benign optimization landscape. Specifically, nonnegative weights allow us to impose acyclicity directly on the adjacency matrix without elementwise squaring of its entries, thus avoiding the well-documented degeneracy of the Karush-Kuhn-Tucker conditions. Computationally, we develop a method of multipliers' algorithm that accommodates both homoscedastic and heteroscedastic noise profiles. Within each iteration, we use block successive convex approximation to minimize the augmented Lagrangian, alternating between proximal gradient steps for the adjacency matrix and closed-form updates for the noise scales. Simulated experiments demonstrate NoCo's improved structural and edge-weight recovery relative to competing methods in a variety of settings, highlighting the benefits of exploiting nonnegativity along with noise adaptivity.
Related
- Concomitant DAG Learning: On the Roles of Noise Adaptivity, Sparsity, and Non-negativity
- Time-warping estimation via stationarity-based learning of the de-warped signal
- The Geometry of Multitask Learning: Sharp Rates for Shared Subspace Estimation
- Parameter estimation in Conditional Sequential Monte Carlo algorithms through Particle Learning
- Agnostic Model-Assisted Estimation with Machine Learning for Survey Data
BibTeX
@misc{navarro2026nonnegative,
title = {{Nonnegative DAG Learning via Concomitant Estimation}},
author = {Madeline Navarro and Gonzalo Mateos and Samuel Rey},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.34595},
eprint = {2609.34595},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.34595},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Nonnegative DAG Learning via Concomitant Estimation
AU - Madeline Navarro
AU - Gonzalo Mateos
AU - Samuel Rey
PY - 2026
DA - 2026/09/28
PB - arXiv
DO - 10.48550/arxiv.2609.34595
UR - https://arxiv.org/abs/2609.34595
AB - We study the problem of learning directed acyclic graphs (DAGs) with nonnegative edge weights from observational data. We propose the Nonnegative and Concomitant (NoCo) DAG estimator, which jointly recovers the weighted graph structure and the exogenous noise variances in the linear structural equation model for the observations. Different from prior art, this noise-adaptive formulation blends a smoothed concomitant lasso criterion with a simpler log-determinant acyclicity constraint that exploits nonnegativity and yields a more benign optimization landscape. Specifically, nonnegative weights allow us to impose acyclicity directly on the adjacency matrix without elementwise squaring of its entries, thus avoiding the well-documented degeneracy of the Karush-Kuhn-Tucker conditions. Computationally, we develop a method of multipliers' algorithm that accommodates both homoscedastic and heteroscedastic noise profiles. Within each iteration, we use block successive convex approximation to minimize the augmented Lagrangian, alternating between proximal gradient steps for the adjacency matrix and closed-form updates for the noise scales. Simulated experiments demonstrate NoCo's improved structural and edge-weight recovery relative to competing methods in a variety of settings, highlighting the benefits of exploiting nonnegativity along with noise adaptivity.
ER -