Differentiable Optimization Layers for Guaranteed Fairness in Deep Learning
Abstract
Differentiable optimization layers are traditionally integrated in predict-then-optimize frameworks where a neural model estimates parameters that subsequently serve as fixed inputs to downstream decision-making optimization problems. In this work, we introduce the concept of a "fairness layer": a differentiable optimization layer appended to a model's output layer that guarantees a chosen notion of output parity is satisfied when integrated into a neural network. Additionally, we introduce an online primal-dual inference algorithm that provides provable aggregate fairness guarantees for streaming predictions with arbitrarily small batch sizes, where traditional per-batch constraints become overly restrictive. Numerical experiments demonstrate the effectiveness of the fairness layer and associated algorithm, and theoretical analysis characterizes the layer's differentiability and stability properties during model training and backpropagation. Our code for these experiments is publicly available on GitHub (https://github.com/dtroxell19/FairDL-ICML-2026.git) and our public Python package documentation can be found online: https://dtroxell19.github.io/fairness_training/.
Related
- lapanda: A Matrix-Free Differentiable Solver for Nonconvex Constrained Optimization Layers
- Movable Antenna Enabled Anti-Jamming: From Local Optimality Guaranteed Design to a Unified Antenna Position Optimization Framework
- Planning to Learn: Real-Time Robust Adaptive Control on the GPU with Guaranteed Future Learning
- The Role of Uncertainty in Assessing the Fairness of Machine Learning Models
- Differentiable Synthesis and Yield Optimization of Cross-Coupled Resonator Filters
BibTeX
@misc{troxell2026differentiable,
title = {{Differentiable Optimization Layers for Guaranteed Fairness in Deep Learning}},
author = {David Troxell and Noah Roemer and Guido Montúfar},
howpublished = {arXiv},
year = {2026},
month = may,
doi = {10.48550/arxiv.2605.17118},
eprint = {2605.17118},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.17118},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Differentiable Optimization Layers for Guaranteed Fairness in Deep Learning
AU - David Troxell
AU - Noah Roemer
AU - Guido Montúfar
PY - 2026
DA - 2026/05/16
PB - arXiv
DO - 10.48550/arxiv.2605.17118
UR - https://arxiv.org/abs/2605.17118
AB - Differentiable optimization layers are traditionally integrated in predict-then-optimize frameworks where a neural model estimates parameters that subsequently serve as fixed inputs to downstream decision-making optimization problems. In this work, we introduce the concept of a "fairness layer": a differentiable optimization layer appended to a model's output layer that guarantees a chosen notion of output parity is satisfied when integrated into a neural network. Additionally, we introduce an online primal-dual inference algorithm that provides provable aggregate fairness guarantees for streaming predictions with arbitrarily small batch sizes, where traditional per-batch constraints become overly restrictive. Numerical experiments demonstrate the effectiveness of the fairness layer and associated algorithm, and theoretical analysis characterizes the layer's differentiability and stability properties during model training and backpropagation. Our code for these experiments is publicly available on GitHub (https://github.com/dtroxell19/FairDL-ICML-2026.git) and our public Python package documentation can be found online: https://dtroxell19.github.io/fairness_training/.
ER -