Learning Quantum Systems through Physics-Inspired Artificial Intelligence

Speakers: Bo-Han Wu (UHM)

Quantum technologies, including sensing, communication, and computing, are reshaping modern science by exploiting nonclassical resources such as coherence, entanglement, and squeezing. Despite significant progress, their advancement remains limited by the lack of system-level co-optimization that connects physical hardware and algorithmic design. Conventionally, quantum hardware and information processing have been developed independently, which reduces robustness against noise and resource inefficiency. Machine learning (ML) provides a transformative pathway toward physics-informed co-design, enabling real-time adaptive optimization of complex optical and quantum systems. By embedding the governing physical principles into learning architectures, ML can enhance signal extraction, suppress noise, and dynamically tune device performance directly at the physical layer. In this talk, I will review three representative directions in quantum photonics: quantum radar, quantum repeater, and cluster-state generation, which illustrate the unifying role of photonics in quantum sensing, communication, and computing. I will then present my recent works on the Microring Perceptron (MiRP) and the Bidirectional Nonlinear Optical Tomography (BNOT) method. Both MiRP and BNOT employ classical optical systems assisted by ML for noise-robust signal processing and unbiased device characterization. Looking ahead, extending these ML-driven frameworks into quantum optical domains will enable the discovery of new design principles, integration strategies, and optimization methods that advance the development of scalable and deployable quantum infrastructures.

https://indico.phys.hawaii.edu/event/2717/