<ul><li>Quantum-classical Hybrid Machine Learning (QHML) models are known for their robust performance in anti-cancer drug response prediction, especially with limited datasets.</li><li>Hybrid models are sensitive to data encoding, with suboptimal choices causing stability issues.</li><li>To improve, a novel normalization strategy using a moderated gradient version of $	anh$ was proposed to enhance model stability by transforming neural network outputs without concentrating them at extreme value ranges.</li><li>Evaluation on gene expression and drug response data showed QHML outperforming classical models when data was optimally normalized, signaling potential for quantum computing in biomedical data analysis.</li></ul>

Optimal normalization in quantum-classical hybrid models for anti-cancer drug response prediction

Discover more