Phase Transition in Binary Compressed Sensing via Annealing with Adaptive Regularization A quantum, classical hybrid approach now outperforms simulated annealing across portions of its operating range, achieving lower mean squared reconstruction errors on comparable problems. This improvement stems from predicting optimal regularization parameters using random forest regression trained upon prior search data. Consequently, previously computed solutions can be reused to streamline reconstructions without repeated parameter selection.
Regression Finds Parameters That Reduce Error by up to 15% during Data Reconstruction
Phase Transition in Binary Compressed Sensing via Annealing with Adaptive Regularization A quantum, classical hybrid approach now outperforms simulated annealing across portions of its operating range, achieving lower mean squared reconstruction errors on comparable problems. This improvement stems from predicting optimal regularization parameters using random forest regression trained upon prior search data. Consequently, previously computed solutions can be reused to streamline reconstructi...