The choice of measurement can determine if a quantum computer learns

The choice of measurement can determine if a quantum computer learns
Image: Intellect (generated)

In short: Researchers who modeled variational quantum algorithms found that the way we measure success determines whether a quantum computer can actually learn. By analyzing how different cost functions affect gradients, they discovered that using global measurements can cause training to stall entirely, even in relatively simple circuits.

This matters because many proposed quantum algorithms rely on 'global' observables—properties of the entire system at once. The study shows that these choices can lead to "barren plateaus," where gradients vanish exponentially, making it nearly impossible for an optimizer to find the best settings. In contrast, using local observables keeps the gradient manageable.

The researchers used mathematical proofs and large-scale simulations of up to 100 qubits in a quantum autoencoder to demonstrate these effects. A limitation is that while the study provides rigorous proof for specific circuit types like alternating layered ansatze, it may not apply to all possible circuit architectures.

Why it matters: To make quantum computers trainable, researchers must prioritize 'local' measurements over global ones to avoid mathematical dead ends.

Opens the Intellect+ feed with this card first. Swipe up for the next one.

Sources

To cite this card, link to https://intellectplus.co.uk/app/technology-research-the-choice-of-measurement-can-determine-if-a-quantum-compute.

How we source, check and correct cards: Editorial standards.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

Finding something worth knowing…