Search papers, labs, and topics across Lattice.
This paper explores Gaussian process bandit optimization using quantum kernels, focusing on the balance between expressivity and learnability in the context of near-term quantum computing applications. The authors introduce projected quantum kernels and classical kernel approximation techniques to mitigate the challenges posed by high-dimensional kernels, which can lead to increased cumulative regret. Their empirical results demonstrate that these methods significantly enhance sample efficiency and reduce computational overhead compared to traditional quantum kernels, facilitating scalable optimization in quantum-native tasks.
Projected quantum kernels can drastically improve sample efficiency in Gaussian process optimization, outperforming traditional high-dimensional quantum kernels.
We investigate Gaussian process (GP) bandit optimization with quantum kernels, assuming the mean reward function lies in the reproducing kernel Hilbert space (RKHS) induced by the quantum kernel. This setting is motivated by NISQ-era tasks such as quantum control, state preparation and variational quantum algorithms. While quantum kernels can offer a `quantum advantage' via domain-specific inductive biases, na茂vely using full, high-dimensional kernels increases model complexity and information gain, leading to higher cumulative regret and poor learnability. To address this, we propose projected quantum kernels and classical kernel approximation techniques that reduce feature dimensionality while preserving key quantum properties. Using these approximate kernels, we develop misspecified GP bandit algorithms and derive regret bounds that characterize the trade-off between approximation error and information gain. The regret bounds provide principled guidance for selecting the optimal model complexity. Empirically, our methods outperform full quantum kernels in sample efficiency, while substantially reducing computational overhead, enabling scalable GP optimization for quantum-native applications.
NUS