Natural Resources Canada | Hyperparameter optimization using quantum genetic algorithms

The Canada Centre for Mapping and Earth Observation has tasked the Quantum AlgoLab with investigating various approaches to fine-tuning hyperparameters for their machine learning models employed in flood prediction. Fine-tuning hyperparameters remains an unresolved challenge, research in this direction is therefore crucial. To address this, three distinct methodologies were examined and compared, covering a spectrum of techniques that integrate quantum or quantum-inspired computing with classical genetic algorithms, already established in the domain of hyperparameter fine-tuning. Comprehensive analyses were done, comparing the efficacy of quantum methods against a conventional genetic algorithm.

The implications of this project go beyond flood prediction, with potential applications in diverse fields utilizing machine learning models. The developed methodologies hold promise for implementation in contexts beyond the realm of flood prediction.