EURO Bulletin Board - Webinars
Next EURO ORML Seminar - October 19, 2026, 16:30 CEST. Prof Angela Zhou: Best of Both Prediction and Decision Worlds: Benefits of Decision-Reweighted Prediction Loss
https://euroorml.euro-online.org/
The EURO OSS on Operational Research and Machine Learning is an online seminar series with the goal to brand the role of Operational Research in Artificial Intelligence. The format is a weekly session that takes place every Monday, 16.30-17.30 (CET). It is 100% online-access, and it has leading speakers from Operational Research, as well as neighbouring areas, that will cover important topics such as explainability, fairness, fraud, privacy, etc. We also have the YOUNG EURO OSS on Operational Research and Machine Learning. In each YOUNG session, three junior academics will show their latest results in this burgeoning area. All talks are recorded, and the videos are uploaded to the website.
To receive news from us, please register to the mailing list using this link
https://forms.gle/YWLb6EPKRQnQert68
You will receive information about our upcoming seminars and the link to participate. Our seminars are on Mondays, 16.30-17.30 CET, and we typically send the link on Monday morning.
October 19, 2026, 16.30 – 17.30 (CET)
Best of Both Prediction and Decision Worlds: Benefits of Decision-Reweighted Prediction Loss
Speaker: Prof Angela Zhou
Assistant Professor in Data Sciences and Operations
University of Southern California, Marshall School of Business
Courtesy appointment in the Thomas Lord Department of Computer Science.
Abstract:
We propose a decision-aware version of predict-then-optimize for contextual linear optimization. We reweigh the prediction error by the decision regret incurred by an (unweighted) pilot estimator of costs to obtain a decision-aware predictor, then optimize with cost predictions from the decision-aware predictor. This method can be motivated as a finite-difference, iterate-independent approximation of the gradients of previously proposed end-to-end learning algorithms; it is also consistent with previously suggested intuition for end-to-end learning. This is computationally easy to implement with readily available reweighted prediction oracles and linear optimization, and can be implemented with convex optimization so long as the prediction error minimization is convex. We demonstrate that this approach can lead to improvements over a “predict-then-optimize” framework for settings with misspecified models, and is competitive with other end-to-end approaches.
Posted on 2026-10-14 by Sarah Fores