A tale of perfect fit and phantom optima: how data-driven models can fail in real-time optimization

arXiv:2608.23885 2026 Training 2 ideas extracted · analyzed Sep 1, 2026

What the math gives to ML

The paper identifies a practically important failure mechanism: low prediction error on historical trajectories does not preserve the economically relevant optimum of a dynamical process. Neural surrogate models can create multiple phantom optima, and stochastic gradient training can move from weights that recover the correct optimum to observationally similar weights with poor decision performance. The strongest transfer to neural networks is decision-oriented training and validation: explicitly test and regularize the surrogate's optimal-control landscape rather than relying only on trajectory fit.

Ideas from this paper

Failed on benchmark 2026

Decision-Oriented Optimum Preservation

Train a neural dynamical surrogate not only to reproduce measured trajectories, but also to reproduce the plant's economically optimal decision and objective value. Add a differentiable decision loss obtained by solving the surrogate's inner optimization problem, and reject models that fit observations while producing extra local optima or a shifted optimum.

Useful8/10
Difficulty6/10
Novelty7/10
Paper: A tale of perfect fit and phantom optima: how data-driven models can fail in real-time optimization arXiv:2608.23885
Failed on benchmark 2026

Phantom-Optimum Audit and Optimizer Drift Monitor

Treat the optimized surrogate and the training trajectory as objects that require a decision-level audit. Use multistart optimization to count phantom optima, and periodically evaluate whether stochastic training has changed the surrogate optimum even when validation prediction error remains nearly constant; stop, roll back, or average checkpoints when decision drift exceeds a threshold.

Useful7/10
Difficulty4/10
Novelty8/10
Paper: A tale of perfect fit and phantom optima: how data-driven models can fail in real-time optimization arXiv:2608.23885