Learning Distributionally Robust First-Order Methods for Convex Optimization
Abstract
We propose a distributionally robust approach to learning hyperparameters (e.g., step sizes and momentum coefficients) for first-order methods in convex optimization. Given a dataset of problem instances, we minimize a Wasserstein distributionally robust version of the performance estimation problem (PEP). Our framework unifies two extremes. As the robustness radius vanishes, we recover classical learning to optimize (L2O), and as it grows we recover the PEP worst-case algorithm design problem. We solve the resulting problem with stochastic gradient descent, differentiating through the solution of an inner semidefinite program at each step. We prove high-probability bounds showing that the true risk of the learned algorithm is at most the in-sample L2O optimum plus a slack that shrinks with the sample size, and is no worse than the worst-case PEP bound. On logistic regression, LASSO, and linear programming examples, our learned algorithms achieve strong out-of-sample performance with certifiable robustness, outperforming both worst-case optimal and vanilla L2O baselines.
Related
- First-Order Methods for Wasserstein Distributionally Robust Constrained Optimization
- Data-driven joint optimization of maintenance and spare parts provisioning: A distributionally robust approach
- On tractability, complexity, and mixed-integer convex programming representability of distributionally favorable optimization
- Conformal-DRO: Distributionally Robust Optimization with Conformalized Ambiguity Set
- Calibrating Ambiguity Set via Diagnostic Transport for Distributionally Robust Optimization
BibTeX
@misc{ranjan2026learning,
title = {{Learning Distributionally Robust First-Order Methods for Convex Optimization}},
author = {Vinit Ranjan and Jisun Park and Bartolomeo Stellato},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2605.06585},
eprint = {2605.06585},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.06585},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Learning Distributionally Robust First-Order Methods for Convex Optimization
AU - Vinit Ranjan
AU - Jisun Park
AU - Bartolomeo Stellato
PY - 2026
DA - 2026/09/29
PB - arXiv
DO - 10.48550/arxiv.2605.06585
UR - https://arxiv.org/abs/2605.06585
AB - We propose a distributionally robust approach to learning hyperparameters (e.g., step sizes and momentum coefficients) for first-order methods in convex optimization. Given a dataset of problem instances, we minimize a Wasserstein distributionally robust version of the performance estimation problem (PEP). Our framework unifies two extremes. As the robustness radius vanishes, we recover classical learning to optimize (L2O), and as it grows we recover the PEP worst-case algorithm design problem. We solve the resulting problem with stochastic gradient descent, differentiating through the solution of an inner semidefinite program at each step. We prove high-probability bounds showing that the true risk of the learned algorithm is at most the in-sample L2O optimum plus a slack that shrinks with the sample size, and is no worse than the worst-case PEP bound. On logistic regression, LASSO, and linear programming examples, our learned algorithms achieve strong out-of-sample performance with certifiable robustness, outperforming both worst-case optimal and vanilla L2O baselines.
ER -