33 citations
2 papers
cs.LG2025★ 1 cited
Binarizing Physics-Inspired GNNs for Combinatorial Optimization
Martin Krutský, Gustav Šír, Vyacheslav Kungurtsev +1
Physics-inspired graph neural networks (PI-GNNs) have been utilized as an efficient unsupervised framework for relaxing combinatorial optimization problems encoded through a specif…
astro-ph.SR2016★ 33 cited
A new method to quantify and reduce projection error in whole-solar-active-region parameters measured from vector magnetograms
David A. Falconer, Sanjiv K. Tiwari, Ronald L. Moore +1
Projection error limits the use of vector magnetograms of active regions (ARs) far from disk center. In this Letter, for ARs observed up to 60o from disk center, we demonstrate a m…