[Abstract]
The Mermin-Peres Magic Square is a well-known example that illustrates quantum entanglement. I have written related articles in the past (in Japanese, available [here]). This time, I aimed to deepen my understanding by visualizing the phenomenon, using four different methods:
(1)Display of a disk shape using my own simulator, (2)3D Qsphere representation with Qiskit, (3)Pauli correlation analysis, and (4)Measurement on IBM Quantum real hardware.
🔴Example: Tiny Mermin-Peres Magic
An explanation of this example is provided [here]. As shown in Fig. 1, Alice and Bob each possess two qubits. In the first half, their qubit pairs are prepared in an entangled quantum state, ǀψ₁⟩. After that, there is no further interaction between them. Nevertheless, when the final state ǀψ₂⟩ is measured, a strong correlation is observed between their outcomes (as shown later in Fig. 2 and Fig. 3). Specifically, the number of "1"s in the 2-bit classical measurement results is always even for one party and odd for the other. This reveals a strong inverse (or anti-) correlation.
🔴Confirmation Using a Homemade Quantum Circuit Simulator
Figure 2 shows the results of verifying the behavior of this example using a homemade quantum circuit simulator. The probability and phase of each basis state are displayed on a disk. In addition, numerical lists of the probability amplitudes, probabilities, and phases are provided.
For example, in the quantum state ǀψ₁⟩, if Alice's qubits are in the state ǀ00⟩, Bob's qubits will definitely be in the state ǀ11⟩. In the final state ǀψ₂⟩, if Alice is in the state ǀ00⟩, Bob's state will be either ǀ01⟩ or ǀ10⟩. It is also possible to confirm whether the total number of "1"s in the measured classical bitstrings is even or odd, as described earlier. In other words, this disk representation allows us to visually understand the quantum entangled state shared between Alice and Bob.
🔴3D display of quantum states using Qiskit's Qsphere
Quantum entanglement cannot be separated into individual qubit states, so quantum states cannot be displayed individually on a Bloch sphere. However, by using Qiskit's Qsphere, it is possible to simultaneously display all possible basis states on a sphere, as shown in Fig. 3. Here, the size of the circle at the tip of the quantum state vector is the probability value, and the color indicates the phase. It matches well with the disk display in Fig. 2.
🔴Indication of the strength of quantum entanglement by Pauli correlation
One way to investigate the quantum entangled state in more detail is the Pauli correlation measurement. This is based on the fact that, for example, the expectation value when the quantum state ǀψ⟩ is measured on the Z axis is calculated as follows: ⟨ZZ⟩ = ⟨ψǀ Z⊗Z ǀψ⟩
As an example, for the Bell state ǀψ⟩ = (ǀ00⟩+ǀ11⟩)/√2, calculations show that ⟨ZZ⟩ = 1, which means that the Z measurements of the two qubits are correlated in exactly the same direction. On the other hand, for the simple tensor product ǀψ⟩ = ǀ01⟩, ⟨ZZ⟩ = -1, which shows a strong correlation in the opposite direction but is not entangled.
In Fig. 4, in addition to the Z measurement, X and Y measurements are also performed. Although the above Pauli measurements can calculate correlations, they do not necessarily fully reflect the "quantum entanglement". Therefore, in some cases, measurements other than the Z measurement may be performed.
In Fig. 4, ZZ(q2,q3) on the horizontal axis is the ⟨ZZ⟩ calculation for Bob's quantum bit, and ZZ(q0,q1) on the vertical axis is the ⟨ZZ⟩ calculation for Alice's quantum bit. The color of the square at the intersection represents the product of their values (expectation values). The darker the red, the closer it is to +1.
In the state ǀψ1⟩, if the Z measurement results of Alice's two qubits are in the same direction, then the Z measurement results of Bob's two qubits will also be in the same direction. If, on the other hand, the Z measurement results of Alice's two qubits are in opposite directions, then the Z measurement results of Bob's two qubits will also be in opposite directions. Such a strong correlation is also observed in ⟨XX⟩ and ⟨YY⟩, so they can be said to be in a fully entangled state.
On the other hand, in the state ǀψ2⟩, the Z measurement of the two qubits will have the same direction for either Alice or Bob, but different directions for the other. Therefore, it makes sense that the squares in the right panel of Fig. 4 are dark blue (the product of the expectations is -1).
🔴Measurement results on an actual IBM Quantum computer
Finally, Fig. 5 shows the measurement results on an actual IBM Quantum computer. This is the result of 10,000 shots performed on one of the most advanced machines, ibm_torino (Heron r1). Although there are some errors (likely due to noise) in both the measurements for the quantum states ǀψ1⟩ and ǀψ2⟩, the results well support the calculation results shown in Fig. 2, Fig. 3, and Fig. 4. It is once again amazing how far quantum computers have progressed!