Further Reading: Simulation and Noise Models

Tagged Tier 1 (confident it exists and recommended) and Tier 2 (real and worth seeking, but verify the current version or URL).

The primary sources

• The qiskit-aer documentation on simulation methods. Each method's algorithm, memory scaling, and options. The matrix_product_state page documents the bond-dimension and truncation-threshold parameters that §11.5 only gestures at, and they are the knobs that decide whether MPS succeeds on a borderline circuit. Tier 1.
• The qiskit_aer.noise documentation. Every channel — depolarizing, readout, amplitude and phase damping, thermal relaxation, Pauli, Kraus — plus NoiseModel.from_backend. The reference for §11.7. Tier 1.
• The qiskit.quantum_info documentation on DensityMatrix, state_fidelity, and the entropy measures. The tools for §11.3, and for Chapter 4's entanglement tests. Tier 1.

On stabilizer simulation

• Daniel Gottesman, "The Heisenberg Representation of Quantum Computers" (1998), arXiv:quant-ph/9807006. The theorem. Short, and more approachable than its reputation. Tier 1.
• Scott Aaronson and Daniel Gottesman, "Improved Simulation of Stabilizer Circuits" (2004), Physical Review A 70, 052328. The CHP algorithm — essentially what Aer's stabilizer method implements. Read it to understand how a thousand qubits is possible. Tier 1.
• Bravyi and Gosset, "Improved Classical Simulation of Quantum Circuits Dominated by Clifford Gates" (2016), Physical Review Letters 116, 250501. The basis of extended_stabilizer: how to handle a few $T$ gates, and why the cost grows exponentially in their number. Directly explains §11.4's measurement that extended_stabilizer is approximate. Tier 1.
• Anything on magic state distillation and T-count optimization. The complement: $T$ gates are what break classical simulability and what is expensive to implement fault-tolerantly, which is why Chapter 23's resource estimates are quoted in T-count. Tier 2.

On tensor networks and MPS

• Román Orús, "A Practical Introduction to Tensor Networks" (2014), Annals of Physics 349, 117. The standard readable introduction: matrix product states, bond dimension, entanglement entropy, and why area laws make some states cheap. Written for physicists but accessible. Tier 1.
• Guifré Vidal, "Efficient Classical Simulation of Slightly Entangled Quantum Computations" (2003), Physical Review Letters 91, 147902. The paper that established the connection §11.5 rests on: entanglement, not qubit count, is what makes classical simulation hard. Tier 1.
• The literature on classical simulation of quantum supremacy experiments. Search for tensor network contraction results responding to random-circuit-sampling claims. This is the "classical counterattack" pattern from Chapter 1 §1.5, and reading one exchange end to end teaches the field's epistemics better than any summary. Tier 2 — an ongoing conversation.

On noise models and channels

• Nielsen and Chuang, Chapter 8. Quantum operations and noise channels: the Kraus formalism, depolarizing, amplitude damping, phase damping, and the distinction between them that §11.7 measures. §8.3's treatment of amplitude versus phase damping is exactly Case Study 1's Cases B and D, done properly. Tier 1.
• The Qiskit Experiments documentation on $T_1$, $T_2$, and randomized benchmarking measurements. How the parameters that go into a noise model are measured in the first place. The natural companion to this chapter, and the foundation of Chapter 30. Tier 1.
• Any careful treatment of non-Markovian noise. §11.6's honest limitation: from_backend assumes independent, memoryless errors, and real devices exhibit correlated and time-dependent noise. Knowing what the model omits is part of using it. Tier 2.

On what the computational basis cannot see

The theme this chapter reinforces for the fifth time.

• Chapter 4 Case Study 2 of this book — the entanglement witness, which is the instrument Case Study 1's Case D needs. Tier 1.
• Any treatment of quantum state tomography. The complete answer to "what state is this really" — and its exponential cost, which is why witnesses exist. Chapter 30 §30.6. Tier 1.

Forward references

• Chapter 12 turns §11.7's signature table into a decision procedure for "is this noise or a bug?"
• Chapter 13 attacks the channels measured here — readout mitigation first, because §11.7 shows why it has the best return.
• Chapter 25 relies on stabilizer simulation to run real codes on hundreds of qubits.
• Appendix G collects the noise parameters: what $T_1$, $T_2$, gate error, and readout error mean, and typical values.

Where to go next. If one thing: Vidal's 2003 paper. It is the result that reframes the exponential wall from "qubit count is the enemy" to "entanglement is the enemy," which is both more accurate and far more useful when you are deciding whether your circuit is doing anything a laptop could not.

Then Chapter 12 — real hardware, and the decision procedure this whole part has been building toward.