Bridging Quantum and Classical Computing in Drug Design: Architecture Principles for Improved Molecule Generation
This work provides empirically-grounded architectural guidelines for hybrid quantum-classical models, enabling more effective integration of current quantum computers into pharmaceutical research pipelines, addressing a domain-specific problem in drug discovery.
The paper tackled the problem of unclear optimal model architectures for hybrid quantum-classical machine learning in drug discovery by systematically optimizing the quantum-classical bridge architecture of GANs using multi-objective Bayesian optimization. The result was an optimized model (BO-QGAN) that achieved a 2.27-fold higher Drug Candidate Score than prior quantum-hybrid benchmarks and a 2.21-fold higher score than the classical baseline, while reducing parameter count by over 60%.
Hybrid quantum-classical machine learning offers a path to leverage noisy intermediate-scale quantum (NISQ) devices for drug discovery, but optimal model architectures remain unclear. We systematically optimize the quantum-classical bridge architecture of generative adversarial networks (GANs) for molecule discovery using multi-objective Bayesian optimization. Our optimized model (BO-QGAN) significantly improves performance, achieving a 2.27-fold higher Drug Candidate Score (DCS) than prior quantum-hybrid benchmarks and 2.21-fold higher than the classical baseline, while reducing parameter count by more than 60%. Key findings favor layering multiple (3-4) shallow (4-8 qubit) quantum circuits sequentially, while classical architecture shows less sensitivity above a minimum capacity. This work provides the first empirically-grounded architectural guidelines for hybrid models, enabling more effective integration of current quantum computers into pharmaceutical research pipelines.