Pauli Noise, Twirling & Error Models
Turning arbitrary noise into stochastic Pauli errors.
Real noise is messy: coherent over-rotations, correlated crosstalk, slow drifts. Yet almost every threshold calculation assumes random Pauli errors. The bridge is twirling: randomize your gates and any noise channel becomes a Pauli channel with the same average fidelity. Along the way you will meet the right, and the misleading, ways to measure how bad an error is.
What you’ll learn
- Represent a channel by its Pauli matrix, and convert between Pauli error rates and Pauli fidelities.
- Prove that Pauli twirling projects any channel onto the Pauli channel with the same diagonal .
- Prove .
- Define the diamond distance, explain why it rather than average infidelity governs fault tolerance, and compare coherent with stochastic errors.
Channels in the Pauli basis
Expand every Kraus operator of an -qubit channel in the Pauli basis (phases dropped, , ). Writing defines the process matrix :
Every non-identity Pauli anticommutes with exactly half of :
Multiply by , sum over , and use the bilinearity :
So the fidelities and the error rates determine each other. In particular, the probability of no error is the average Pauli fidelity. This is how modern Pauli-noise learning works: the are measured as decay rates of Pauli expectation values and transformed into error rates.
Twirling
In practice you twirl a noisy gate, not bare noise. For a Clifford gate , insert a random Pauli before it and the Pauli after it. Ideally the two cancel, while the noise in between is twirled. Compiling the random Paulis into neighboring single-qubit gates costs no extra depth: this is randomized compiling (Wallman & Emerson 2016), building on Knill's Pauli-frame randomization.
| Pauli before the CNOT | Pauli after it that undoes the first |
|---|---|
A gate suffers a coherent over-rotation, . What is its Pauli twirl?
- The identity channel: twirling removes the error
- The depolarizing channel with
- A dephasing channel: errors with probability
- A bit-flip channel with probability
Show the answer
C. A dephasing channel: errors with probability
The single Kraus operator is , so and , plus off-diagonal terms. Twirling keeps the diagonal: .
Process fidelity and average fidelity
Two fidelities quantify how close a channel is to the identity. The entanglement (process) fidelity sends half of a maximally entangled state through the channel; the average fidelity averages over pure inputs drawn from the unitarily invariant (Haar) measure:
Write the integrand with two copies of :
The Haar average of commutes with every and is supported on the symmetric subspace, so it is the normalized projector onto it ( is the swap operator):
Use and :
Completeness gives , and :
Because twirling preserves , it preserves both fidelities. Twirling over the Clifford group, a unitary 2-design, goes further and yields the depolarizing channel with the same ; its decay under random Clifford sequences is what randomized benchmarking measures. For a qubit the average error rate is , so a Pauli channel has .
A single-qubit Pauli channel has . What is its average gate infidelity , in units of ?
Show the answer
2 × 10⁻³
, so . The average fidelity counts inputs that an error leaves unharmed: an error does nothing to , for example.
The diamond norm, and why coherent errors are dangerous
Unlike average infidelity, the diamond distance is a worst-case quantity that composes well: , and it is stable under tensoring with the identity. That is why threshold theorems are proven for diamond-norm error rates. How do the two measures compare?
Lower bound. Feed in and test with the projector , which detects any deviation with probability :
Upper bound. The Fuchs–van de Graaf inequality bounds the distance between Choi states by , and the diamond norm exceeds the trace norm of the normalized Choi operator by at most a factor :
Pauli channels saturate the lower bound. For a qubit this means .
Coherent errors sit near the square-root end. For , but . With that is :
| Error model | What can fail | Surface-code threshold |
|---|---|---|
| Code capacity | data qubits only, e.g. independent depolarizing errors | of order 10% |
| Phenomenological | data qubits and syndrome bits | a few % |
| Circuit-level | every gate, idle step, preparation and measurement | about 1% or below |
| Erasure | errors at known locations | 50% with perfect checks (bond percolation) |
Key takeaways
- Channels expand in the Pauli basis as ; Pauli channels have diagonal , and their error rates and Pauli fidelities are related by a Walsh–Hadamard transform.
- Pauli twirling (in practice, randomized compiling) deletes the off-diagonal , turning any noise into stochastic Pauli errors with the same fidelities.
- , proven with the Haar average over the symmetric subspace.
- The diamond distance is the worst-case, composable error measure: , with Pauli noise at the bottom and coherent errors near the top.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
Twirling a channel over the -qubit Pauli group preserves which of the following?
- The full matrix
- The diamond distance to the identity
- The diagonal of , and hence both and
- Only the purity of output states
Show the answer
C. The diagonal of , and hence both and
The twirl keeps and deletes for . Since and , both fidelities are unchanged, while the diamond distance generally drops.
A two-qubit gate has process fidelity . What is its average gate fidelity?
Show the answer
C.
: .
Which statement about a single-qubit Pauli channel with total error probability is correct?
- Its diamond distance to the identity is , and its average infidelity is
- Its diamond distance to the identity is
- Its average infidelity is
- Its diamond distance to the identity is
Show the answer
A. Its diamond distance to the identity is , and its average infidelity is
Pauli channels saturate the lower bound: , while .
Why are fault-tolerance thresholds stated in terms of the diamond norm rather than average infidelity?
- Because the diamond norm is easier to measure experimentally
- Because average infidelity can be negative
- Because the diamond distance is always smaller, which gives better thresholds
- Because it bounds the worst case over all inputs, including entangled ones, and adds up under composition
Show the answer
D. Because it bounds the worst case over all inputs, including entangled ones, and adds up under composition
In a long computation, errors act on states entangled with the rest of the computer. The diamond distance bounds the damage for every such input and is subadditive under composition, so a per-gate bound becomes a whole-circuit bound.
Further reading
- C. H. Bennett, D. P. DiVincenzo, J. A. Smolin & W. K. Wootters, "Mixed-state entanglement and quantum error correction", Physical Review A 54, 3824 (1996).
- M. A. Nielsen, "A simple formula for the average gate fidelity of a quantum dynamical operation", Physics Letters A 303, 249 (2002).
- J. J. Wallman & J. Emerson, "Noise tailoring for scalable quantum computation via randomized compiling", Physical Review A 94, 052325 (2016).
- A. Yu. Kitaev, "Quantum computations: algorithms and error correction", Russian Mathematical Surveys 52, 1191 (1997).
- J. Watrous, The Theory of Quantum Information, Ch. 3 (Cambridge, 2018).