Professor Junhua Zhao and his team at The Chinese University of Hong Kong, Shenzhen (CUHK-Shenzhen), are at the forefront of exploring the synergistic application of quantum computing and large-scale models. Early in his career, drawing upon his computer science expertise, Zhao Junhua applied quantum computing to tackle intricate optimization challenges within power systems. In recent times, he has observed the issue of ensuring reliable long-range inference when deploying large models in the power sector. To address this, he introduced a novel method: reframing the selection of large model inference paths as a combinatorial optimization problem that can be resolved using quantum devices.

Building on this concept, the team developed the QART (Quantum-Augmented Reasoning Transformer) architecture, which eliminates the need for re-pretraining the foundational model. Initial tests have demonstrated enhanced inference performance across a majority of benchmark tasks. Moreover, Zhao Junhua established QuantumMind, a quantum AI startup committed to driving the commercialization of relevant technologies. The company targets high-stakes, highly regulated sectors such as power and finance, where decision-making errors can result in substantial costs. It leverages specialized quantum computing devices and adopts a hybrid architecture that integrates CPUs, GPUs, and QPUs.

At present, the company's primary focus is on validating the technology's value in real-world business environments, promoting the deployment of industry-specific agents, and exploring quantum-assisted model self-iteration. Nevertheless, the current path to commercialization still grapples with cost challenges. Widespread adoption is anticipated only when quantum computing costs become acceptable to the industry.