The Signal
Ott and Colleagues Use Logical Qubits to Find Rare Signals in Noise
A Physical Review Letters paper by Robert Ott and colleagues proposes quantum error correction for detecting weak, randomly timed signals amid local noise. Its analysis and simulations find an advantage over conventional sensing in specifie
Quantum.world Team · Written with AI · fact-checked ·
“For rare signals occurring at random times in the presence of local Markovian noise, we explicitly demonstrate an improved sensitivity of our approach over more conventional sensing strategies.”
Robert Ott, Torsten V. Zache, Soonwon Choi, Adam M. Kaufman and Hannes Pichler, “Rare-Event Quantum Sensing Using Logical Qubits,” Physical Review Letters
A code turns pulse patterns into a logical phase
The paper, published online by the American Physical Society on October 30, 2026, proposes using quantum error correction (QEC) to distinguish rare signals from noise. The authors model weak pulses that arrive at unknown times according to a Poisson process, with a known average rate. Their target is to estimate each pulse’s strength, represented by a small rotation angle, while local Markovian noise disrupts the sensors.
The approach uses repeated measurements of a quantum code’s stabilizers—checks that reveal errors without directly reading the encoded information. Those measurements do more than flag faults: their nonlinear processing responds differently to a collective signal pulse and to local noise. That distinction lets the protocol preserve a signal record even when the signal and noise affect the same physical degrees of freedom, a setting where standard QEC-based sensing faces a known limit.
The seven-qubit Steane code supplies the test case
Ott and colleagues work through the protocol using the [[7,1,3]] Steane code, which encodes one logical qubit in seven physical qubits and can correct a single-qubit error. In their model, a pulse applies a small rotation across the sensors. Syndrome extraction converts that event into a probabilistic logical operation: different measurement outcomes produce different logical phases, whose average scales with the cube of the pulse angle, ε³.
That cubic scaling is a real cost: the logical signal is weaker than the physical pulse. The trade is longer coherent acquisition. Frequent syndrome checks suppress physical dephasing, while the stochastic outcomes of those checks add their own logical dephasing. For the Steane-code analysis, the paper gives the physical-noise contribution as 42γ²δt, where γ is the physical noise rate and δt the interval between checks; the signal’s random measurement outcomes contribute a term proportional to Rε⁴.
The gain depends on long sensing and rare events
The authors compare the logical protocol with ordinary Ramsey sensing and with classical error detection, which filters some noise by discarding selected measurement outcomes. In the long-time regime and the idealized limit of very frequent syndrome extraction, their calculated Fisher-information advantage over Ramsey sensing approaches γ/R. Against classical error detection, it approaches 1/ε². Those are model-dependent comparisons, not a measured improvement from a sensor experiment.
The result also has boundaries. The protocol targets weak, infrequent pulses, and the paper notes that the expected advantage does not generally extend to high signal rates. Its analysis relies on many pulses during the sensing interval; it also says that when the accumulated phase grows beyond 2π, that phase wrapping must be handled. Entangling several logical sensors can improve sensitivity for shorter sensing times, but the scaling returns to the standard quantum limit at longer times.
Hardware speed sets the practical test
The study presents analytical calculations and numerical simulations, not a demonstration on a physical device. It focuses on pure dephasing in its worked example, while describing extensions to other local Markovian noise when a suitable code can correct those errors. For pure dephasing, the authors say syndrome extraction alone can suffice; other noise models generally call for active error correction.
Putting the proposal to work would require fast, reliable syndrome measurements. The paper identifies pulse duration, the clock speed of the hardware and gate-level noise during syndrome extraction as limits on performance. Those engineering costs determine whether the longer logical coherence can outweigh the signal’s reduction to a third-order phase in a real sensor.