The researchers show that small inaccuracies in quantum processors can linger, evolve and correlate across different computational steps, creating a form of temporal memory that standard error models do not capture.
Lead author Dr Christina Giarmatzi explains that these machines can retain a memory of their own errors, which may be classical or quantum in nature depending on how the faults connect through time.
Many quantum algorithms and protocols assume that hardware noise is Markovian, meaning that each error is independent and does not depend on earlier events, but the study demonstrates that this assumption often fails in practice.
This time-linked behavior is now identified as a central obstacle to building practical, large-scale quantum computers, because conventional error-correction schemes are typically designed for uncorrelated, memoryless noise.
To probe these effects, the team carried out experiments on superconducting quantum processors, using devices operated in the University of Queensland laboratory as well as chips accessed through IBM's cloud-based quantum computing services.
By analyzing how quantum processes evolve at multiple points in time, the researchers were able to reconstruct the full temporal dynamics of the noise, tracking when disturbances arise and how they propagate through successive operations.
According to Dr Giarmatzi, this reconstruction offers a new way to see how quantum systems behave when their errors are correlated across time, providing information that is essential for developing more reliable, error-controlled machines.
The work indicates that even state-of-the-art quantum processors display subtle but important patterns of time-correlated noise, including effects that originate from neighboring qubits interacting on the same chip.[2]
These insights are expected to guide the design of improved characterization tools and error-correction strategies that explicitly account for temporal correlations, an important step on the path to dependable, fault-tolerant quantum computers.
Team member Tyler Jones notes that the project demonstrates how theoretical models of time-correlated errors can be implemented and tested on real hardware, helping to refine both the devices and the methods used to benchmark them, and emphasizes that robust characterization of time correlations is needed to build powerful quantum machines.
The researchers have released their experimental data and code for public use, and the full study appears in the journal Quantum.
Research Report:Multi-time quantum process tomography on a superconducting qubit
Related Links
Macquarie University
Understanding Time and Space
| Subscribe Free To Our Daily Newsletters |
| Subscribe Free To Our Daily Newsletters |
