Reproduce “Self-Supervised Tuning of Low-Fidelity Optimization Models via High-Fidelity Feedback” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Gonzalo E. Constante-Flores, André H. Quisaguano-Paredes, Asha Ramanujam, Rishabh Gupta, Can Li
The method tunes selected parameters of a low-fidelity optimization model with a learned correction map, then solves the model. During training, a high-fidelity evaluation problem supplies feedback, eliminating high-fidelity optimal decision labels but not high-fidelity solves. On a transductive 30,000-bus OPF test, it achieved 0.320% mean relative objective error and 0.478 MW P90 feasibility violation, without establishing global optimality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.