What the study found
Uncorrected local noise can make quantum circuits effectively shallow, even when the circuits are not physically shallow. The authors also report that, for cost functions built from local observables, quantum circuits under any non-unital noise do not exhibit barren plateaus.
Why the authors say this matters
The study suggests that, without fault-tolerant quantum error correction, noisy quantum computers are limited in what they can do. The authors conclude that noisy quantum circuits are unlikely to provide an advantage over shallow circuits for algorithms that estimate observable expectation values, including many variational quantum machine learning proposals, unless the circuits are carefully engineered to use the noise.
What the researchers tested
The researchers studied the impact of uncorrected local noise on logical quantum circuits, motivated by near-term hardware. They analyzed observable expectation value estimation and cost functions composed of local observables, and they considered any circuit architecture.
What worked and what didn't
The authors show that noise effectively truncates most quantum circuits to logarithmic depth for the task of estimating observable expectation values. They also prove that circuits under non-unital noise do not show barren plateaus for local-observable cost functions. At the same time, they design a classical algorithm that can estimate observable expectation values within any constant additive accuracy with high probability over the choice of the circuit.
What to keep in mind
The results are about noisy circuits without successful fault-tolerant quantum error correction, and they focus on observable expectation value estimates. The abstract does not describe further limitations beyond these scope conditions.
Key points
- Uncorrected local noise can make quantum circuits effectively logarithmic in depth for observable expectation value estimation.
- Circuits with any non-unital noise do not exhibit barren plateaus for cost functions made from local observables.
- The authors give a classical algorithm that estimates observable expectation values to any constant additive accuracy with high probability over the circuit choice.
- The paper concludes that noisy quantum circuits are unlikely to beat shallow circuits for many variational quantum machine learning proposals unless designed to use the noise.
Disclosure
- Research title:
- Noisy quantum circuits can behave like shallow circuits
- Publication date:
- 2026-04-02
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.