Robert Huang [Caltech]

Abstract:

Quantum technology has the potential to revolutionize how we learn about the physical world. Quantum AI agents with access to the full suite of quantum technology---quantum sensors, quantum memory, and quantum computers---can achieve significant advantages over classical AI agents using conventional technologies. This talk examines our theoretical understanding of these quantum advantages, which has primarily focused on learning many-body quantum systems. To broaden the impact of quantum AI agents, we must extend these capabilities to more classical sensing and learning tasks. Specifically, I will present our recent work (https://arxiv.org/abs/2501.07625) developing a quantum computing-enhanced protocol for a fundamental and practically important sensing problem: detecting oscillating classical fields with unknown frequency. In the strong field and large bandwidth regime, our approach achieves a substantial beyond-quadratic advantage in sensing time. This work illuminates the potential for quantum AI agents to substantially enhance our learning and sensing capabilities in the classical world.

Research interests: I study the intersection of learning, physics, information, and quantum computation. Current research directions include:

- Can quantum sensors enhanced by quantum computers achieve exponential advantages in sensing classical fields?

- How to develop a theory of fault-tolerant quantum AI agents that combine quantum sensing, processing, and memory?

- What provable quantum advantages can small-scale quantum computers achieve in analyzing vast amounts of classical data?

- What are the fundamental physical limits (energy, time, space) on learning? Is there a difference between classical and quantum physics?

Robert Huang is visiting Stanford from Feb 27 to Apr 11, 2025. Please contact him directly via email (hsinyuan@caltech.edu) if you hope to meet with him.