<![CDATA[Drug discovery increasingly relies on accurate simulation of molecular Hamiltonians, yet classical computational methods face exponential scaling barriers when modeling complex quantum systems. Recent advances in quantum machine learning (QML) and the availability of Noisy Intermediate-Scale Quantum (NISQ) devices offer new opportunities to accelerate molecular simulation despite hardware noise and qubit limitations. This study aims to evaluate the effectiveness of QML-based variational algorithms in improving the efficiency and accuracy of Hamiltonian simulation for drug-relevant molecules on NISQ platforms. A hybrid quantum–classical methodology was employed, combining variational quantum eigensolvers, noise-aware circuit optimization, and supervised learning models trained to predict energy landscapes. Experimental simulations were performed using IBM-Q and Rigetti NISQ architectures, supported by classical benchmarks for validation. The results demonstrate that QML-enhanced variational circuits significantly reduce computational depth while maintaining competitive accuracy compared to classical methods, particularly for medium-sized molecular systems. The findings also reveal that noise-adaptive training improves algorithm robustness, enabling more reliable energy estimation under realistic quantum noise conditions. The study concludes that QML provides a promising pathway for accelerating early-stage drug discovery by enabling efficient molecular Hamiltonian simulation on current-generation quantum hardware. Further integration of error mitigation and scalable QML frameworks will be essential for future advancements.]]>
Copyrights © 2025
