Wyame Benslimane | Ph.D. Candidate, IEOR, UC Berkeley

I am a fifth-year Ph.D. candidate in Industrial Engineering and Operations Research at UC Berkeley, advised by Prof. Paul Grigas. I graduated from École CentraleSupélec in 2021 with an Engineering Degree in Applied Mathematics.

My research focuses on data-driven decision-making under uncertainty, at the intersection of machine learning and optimization. I am particularly interested in how leveraging the structure of optimization problems can improve decision quality, a central motivation behind decision-focused learning.

More broadly, I aim to develop learning and optimization methods that combine theoretical guarantees with practical relevance, accounting for the underlying structure of decision problems and the downstream consequences of the decisions they produce.

My recent work extends decision-focused learning to online settings, where decisions must be made sequentially and updated using partial feedback. I am currently exploring how to extend these methods to settings where satisfying constraints is a key component of decision quality.

News

Nov 02, 2026	I will present Decision-Focused On-Policy Learning for Contextual Linear Optimization with Partial Feedback at the INFORMS Annual Meeting on Monday, November 2.

Selected Publications

arXiv

Decision-Focused On-Policy Learning for Contextual Linear Optimization with Partial Feedback

Wyame Benslimane, Tinghan Ye, Pascal Van Hentenryck, and Paul Grigas

Jun 2026

arXiv Bib

@misc{benslimane2026decisionfocusedonpolicylearningcontextual,
  title = {Decision-Focused On-Policy Learning for Contextual Linear Optimization with Partial Feedback},
  author = {Benslimane, Wyame and Ye, Tinghan and Van Hentenryck, Pascal and Grigas, Paul},
  year = {2026},
  month = jun,
  archiveprefix = {arXiv},
  primaryclass = {cs.LG},
}

NeurIPS

Smart Surrogate Losses for Contextual Stochastic Linear Optimization with Robust Constraints

Hyungki Im, Wyame Benslimane, and Paul Grigas

Advances in Neural Information Processing Systems, Dec 2025

Bib PDF

@article{im2025smart,
  title = {Smart Surrogate Losses for Contextual Stochastic Linear Optimization with Robust Constraints},
  author = {Im, Hyungki and Benslimane, Wyame and Grigas, Paul},
  journal = {Advances in Neural Information Processing Systems},
  volume = {38},
  pages = {75149--75174},
  year = {2025},
  month = dec,
}

CPAIOR

Self-supervised Penalty-Based Learning for Robust Constrained Optimization

Wyame Benslimane and Paul Grigas

In International Conference on the Integration of Constraint Programming, Artificial Intelligence, and Operations Research, May 2025

Bib PDF

@inproceedings{benslimane2025self,
  title = {Self-supervised Penalty-Based Learning for Robust Constrained Optimization},
  author = {Benslimane, Wyame and Grigas, Paul},
  booktitle = {International Conference on the Integration of Constraint Programming, Artificial Intelligence, and Operations Research},
  pages = {103--112},
  year = {2025},
  month = may,
  organization = {Springer},
}