A Solver-Free Training Method for Predict-then-Optimize

arXiv:2606.19587 · stat.ML, cs.LG · Submitted 2026-06-17 · Read on arXiv

Beichen Wan, Mo Liu

stat.ML, cs.LG

Submitted: 2026-06-17

Comments: Accepted by ICML 2026

License: http://creativecommons.org/licenses/by/4.0/

The gist: We propose a scalable method for training prediction (machine learning) models in the predict-then-optimize paradigm, where model outputs serve as coefficients for a subsequent linear optimization

Terminology

Abstract

We propose a scalable method for training prediction (machine learning) models in the predict-then-optimize paradigm, where model outputs serve as coefficients for a subsequent linear optimization task. Directly minimizing the empirical decision regret is intractable for linear programming and combinatorial optimization since the decision mapping is piecewise constant, and the gradients are zero almost everywhere. While existing methods address this by smoothing the differentiation process, they suffer from scalability issues, since a computationally expensive solver call is required for every gradient evaluation. To address this, we propose a decision-focused learning pipeline based on a measure transformation principle, which yields a new surrogate loss that is completely optimization-solver-free during training. We establish theoretical guarantees, including Fisher consistency and excess risk bounds. Empirically, our method achieves decision quality competitive with state-of-the-art methods while reducing training time by orders of magnitude.

Sources

Related papers