Stop trying to eliminate noise in analog quantum computers; it turns out noise actually makes the models learn better.
April 17, 2026
Original Paper
Noise-enhanced quantum kernels on analog quantum computers
arXiv · 2604.12476
The Takeaway
In almost every engineering field, noise is the enemy. This paper shows that in analog quantum kernels, operational noise actually increases the model's complexity and expressivity. It effectively acts as a regularizer or a booster for the machine's 'learning' capacity, allowing it to represent more complex patterns. This changes how we build quantum hardware—instead of chasing 'zero noise,' we should be looking for the 'optimal noise' that maximizes performance. It's a counterintuitive discovery that could save millions in hardware stabilization costs.
From the abstract
The quantum kernel method, a promising quantum machine learning algorithm, possesses substantial potential for demonstrating quantum advantage. Although the majority of the quantum kernel is constructed in the context of gate-based quantum circuits, inspired by the idea of analog quantum computing, here we construct an analog quantum kernel and a hybrid quantum kernel, and show their competitiveness against other kernel methods in a benchmarking task and the practical problem of estimating non-M