Finite Systems in the Space of Projective Measurements

1 Introduction

• The 2^N × 2^N discrete phase-space representation becomes difficult to interpret at larger particle numbers and can obscure information density and its location.The paper identifies this as a complication for analyzing Hamiltonian evolution.Source: §1 · p. 1
• The paper constructs an approximate star-product analog for the Q function of operator products, expressed using the operators’ Q and P functions in collective-measurement space.The operators considered include a density matrix and Hamiltonian.Source: §1 · p. 1
• The study examines both permutation-invariant and noninvariant states and Hamiltonians.The stated scope includes Hamiltonians with either permutation symmetry or no such symmetry.Source: §1 · p. 1
• Rewriting the Liouville equation in the proposed representation provides a way to analyze quantum-state evolution.The introduction frames collective measurements as a route to global system information when measurements of specific particles are generally impossible in the setting considered.Source: §1 · p. 1

2 Discrete Phase Space and Collective Measurements

• Discrete phase-space functions can represent states and Hermitian operators, with this work focusing on the Q and P functions.The passage also names the Wigner function as another common example.Source: §2 · pp. 1–2

2.1 Discrete space

• Finite-field elements are associated with binary strings through a self-dual basis, whose coordinates are obtained by taking traces.The paper states that a_i = tr(αθ_i) for the chosen self-dual basis.Source: §2.1 · p. 2
• An N-qubit system is labeled by binary strings and represented on a 2^N × 2^N discrete phase-space grid.Grid coordinates are pairs of N-component binary strings, and generalized Pauli operators are defined on this space.Source: §2 · p. 1, §2.1 · p. 2
• The discrete coherent-type states are labeled by phase-space coordinates and form a complete set over the discrete phase space.The fiducial state |ξ⟩ is an SU(2) coherent state invariant under particle permutations and is a product of N identical two-level states.Source: §2.1 · p. 2

2.2 Distribution functions in discrete space

• The paper focuses on discrete Q and P functions rather than the Wigner function, whose dynamics involve an arbitrary choice of phase accompanying the displacement operator.The P function is defined through an expansion in coherent states and also through a kernel; the Q function is defined as an expectation value and can likewise be constructed using a kernel.Source: §2.2 · p. 2, §2.2.1 · p. 2, §2.2.2 · p. 2

2.2.2 The Q function

• The Q and P mappings are dual, and the paper gives identities connecting them and the expectation value of an operator.For state distributions, a normalization is also specified.Source: §2.2.2 · p. 2
• Figure 2.1 displays discrete Q functions for a coherent state and a GHZ state, while the paper notes that such displays contain excessive information for collective-property studies.The display problem becomes more pronounced as particle number increases.Source: Figure 2.1 · p. 3, §2.2.2 · p. 3

2.3 Measurements of collective operators

• Collective observables offer global information and a useful statistical description when particles cannot be distinguished and individual-particle measurements are inefficient or impossible.The paper identifies collective measurements as the accessible measurements in its experimental setting.Source: §2.3 · p. 3

2.4 The space of collective measurements

• The projected Q function contains the information needed to compute expectation values of permutation-invariant observables, even for states that are not symmetric.Projection discards some original state information but retains what is required for collective-operator expectations; the projected function can also classify states by localization.Source: §2.4 · p. 3, §2.4.1 · p. 4

2.4.1 States in the space of collective measurements

• The projected Q function maps discrete phase-space coordinates to a three-dimensional collective-measurement space using the Hamming weights of α, β, and α + β.The coordinates m, n, and k are those three Hamming weights, with allowed values constrained by possible particle permutations.Source: §2.4.1 · pp. 3–4
• The projected functions provide a clearer view of information density than the full discrete-grid functions, and are defined on a discrete tetrahedron.In the figure’s representation, sphere size and color encode projected-Q values.Source: §2.4.1 · p. 4

2.5 Quantum states in two-level systems

• The analysis covers permutation-invariant and noninvariant pure states, evolving both with permutation-invariant and noninvariant Hamiltonians.Source: §2.5 · p. 4

2.5.1 Permutation-invariant states

• The invariant-state examples include GHZ and Dicke states, with GHZ spanning the all-ground and all-excited configurations.The Dicke-state definition uses all possible particle permutations.Source: §2.5.1 · pp. 4–5
• The W state is an equal-coefficient superposition with exactly one particle in |1⟩ and is the K = 1 case of a Dicke state.Source: §2.5.1 · p. 5
• For three particles, W and GHZ states are both nonbiseparable, but only the remaining W pair stays entangled after one particle is lost.After one particle is lost, a GHZ state becomes separable.Source: §2.5.1 · p. 5

2.5.2 States that are not permutation-invariant

• The domain-wall state interpolates from all particles in |0⟩ to all in |1⟩, with a magnetic interpretation of moments gradually reorienting along a chain.Source: §2.5.2 · p. 5
• The NOON state superposes configurations with N/2 particles in each level and corresponds in optics to photons split between two modes [20].Source: §2.5.2 · p. 5
• The singlet-type state has zero total angular momentum, with half the particles’ spins in each direction; positronium is given as an example.Source: §2.5.2 · p. 5

• The objective is to study quantum dynamics in measurement space (m, n, k), using the Liouville equation as the starting point.Source: §2.6 · p. 5
• The Liouville equation is rewritten in terms of a Q function to express the system’s dynamics in quasidistribution form.Source: §2.6 · p. 5
• Because the discrete-phase-space star product defines the Q function of an operator product, the method seeks an analog for Q functions projected onto (m, n, k).Source: §2.6 · p. 5

• The product-function construction expands the density matrix and Hamiltonian using Eq. (2.16), evaluates integrals with Cauchy’s formula, and gives a compact kernel form.The kernel is calculated numerically once for each N and can then be used for any state and Hamiltonian.Source: §3.1 · pp. 5–7
• The projection relation is exact for permutation-invariant states, is expected to work well for permutation-invariant Hamiltonians, and may deviate for noninvariant Hamiltonians.Source: §3.1 · p. 6
• For a non-permutation-invariant Hamiltonian, the product construction uses an approximation; exchanging the operator order replaces Ψ with its complex conjugate.Source: §3.1 · p. 7
• The projected star-product analog combines projected Q and P functions to express the projected Q function of an operator product.Source: §3.1 · p. 7
• The evolution equation is put in matrix form over the measurement-space points, with dimension M = (N + 1)(N + 2)(N + 3)/6, and evolved using the matrix eigenvalues and eigenvectors.Source: §3.1 · p. 7

4.1 Hellinger distance

• The Hellinger distance compares approximate and exact projected distributions, with D = 0 indicating coincidence and D = 1 complete distinctness.It is used to quantify their proximity in measurement space.Source: §4.1 · p. 8

4.2 Evolution under different Hamiltonians

• For permutation-invariant state and Hamiltonian examples, approximate projected-Q evolution closely matches exact evolution, including the GHZ and coherent-state cases.The GHZ distributions are described as practically identical; the coherent-state distributions are very similar.Source: Figure 4.1 · p. 8, §4.2 · p. 8, Figure 4.2 · p. 9
• For the domain-wall state under the permutation-invariant LMG Hamiltonian, approximate and exact distributions are very similar, and both state types have Hellinger distances of order 10^-5.The domain-wall distribution evolves from a broad spread into a more defined central concentration.Source: Figure 4.3 · p. 9, §4.2 · p. 9
• Under the LMG Hamiltonian, some noninvariant states show errors as small as order 10^-10, and the coherent-state test gives distances of order 10^-5 for ˜g = 0.5, 1, and 1.5.For Eq. (4.6), distances exceed those under the permutation-invariant example even at short times, and the NOON state performs better than the coherent state.Source: §4.2 · pp. 9, 12, Figure 4.4 · p. 10

5 Conclusions

• An error of order 10^-5 or smaller was obtained for permutation-invariant states and Hamiltonians, measured by Hellinger distance against exact evolution.Noninvariant states also evolved accurately when the Hamiltonian was permutation-invariant.Source: §5 · p. 12
• For noninvariant Hamiltonians, errors are not small for every state at long times, although the approximation performs well at short times for both state classes.Which initial states pair well with a noninvariant Hamiltonian remains unresolved.Source: §5 · p. 12
• Across the LMG transition expected at g̃ = 1, the approximation continued to perform well with very small errors over long evolution times.The transition is controlled by g.Source: §5 · p. 12
• The method performed well in the reported N = 6 calculations, while smaller N increased error, suggesting—but not establishing—improvement for N ≫ 1.With permutation-invariant Hamiltonians it worked over long times for both state classes; at short times it worked for most studied states without requiring Hamiltonian symmetry.Source: §5 · p. 12

5.1 Future work

• Kernel construction requires substantial computation time that increases exponentially with N, despite being needed only once per particle number.A simpler kernel expression could enable larger-particle studies and clearer characterization of the projected Q function.Source: §5.1 · p. 12
• Extending the method to prime-power dimensions p^n with p ≠ 2 is identified as an interesting direction.Source: §5.1 · p. 12

Appendix A: The kernel Ψm,n,k,m′,n′,k′

• The derivation starts from an exact expression for the kernel and expands it.Source: Appendix A · p. 13
• The derivation rewrites the state using the commutation relation and state definition, then applies the operators.Source: Appendix A · p. 13
• The kernel expression includes character and Kronecker-delta factors, with weights defined as sums of nonzero finite-field components in the stated basis.Source: Appendix A · p. 14
• The P kernel is expanded using Eq. (2.11), followed by expansion of an operator expectation-value term.Source: Appendix A · p. 14
• The derivation also uses a Kronecker-delta representation and groups terms with shared index dependencies.Source: Appendix A · p. 14

Appendix B: The polynomial

• Evaluating the sums in the expression for Ψ yields a polynomial.Source: Appendix B · p. 14
• The displayed polynomial includes terms with exponential factors indexed by i, p, and related symbols.Source: Appendix B · p. 15
• Additional displayed terms include exponential factors involving q and k′ indices.Source: Appendix B · p. 15