Revealing the Boundary between Quantum Mechanics and Classical Model by EPR-Steering Inequality

Technical Summary: Revealing the Boundary between Quantum Mechanics and Classical Model by EPR-Steering Inequality

Problem Statement
The Werner state serves as a fundamental benchmark for distinguishing between quantum mechanics and classical models. For the two-qubit Werner state, $\rho_W = V |\Phi^-\rangle\langle\Phi^-| + (1-V)\rho_{noise}$, there are three established critical values ($V_c$) defining the boundaries between different types of quantum nonlocality:

1. Entanglement vs. Separability: $V_c^E = 1/3$.
2. EPR Steering vs. Local-Hidden-State (LHS) Model: $V_c^S = 1/2$.
3. Bell Nonlocality vs. Local-Hidden-Variable (LHV) Model: $V_c^B = 1/KG(3)$ (where $KG(3)$ is the Grothendieck constant).

While the boundary $V_c^E = 1/3$ has been rigorously solved using the positive-partial-transpose (PPT) criterion, and $V_c^B$ has been approached via various Bell inequalities (though the exact value remains an open problem due to the impossibility of exhausting all inequalities), the derivation of $V_c^S = 1/2$ via an inequality approach has remained an open challenge. Previous attempts, such as the $N$-setting linear EPR-steering inequalities proposed by Saunders, Jones, Wiseman, and Pryde (SJWP), utilized geometric symmetries of Platonic solids to define measurement directions. However, these inequalities were limited to $N \leq 10$ settings and yielded a classical bound $C_{10} \approx 0.5236$, which, while close to $0.5$, did not strictly prove $V_c^S = 1/2$.

Methodology
The authors propose and analyze optimal $N$-setting linear EPR-steering inequalities. The core methodology involves:

1. Optimization of Measurement Directions: Unlike SJWP's approach which relied on fixed geometric symmetries, this work seeks the optimal set of $N$ measurement directions $\{\vec{b}_j\}$ for Bob that minimizes the classical (LHS) bound $C_N$ for a fixed $N$.
2. Numerical Simulation: The authors employ a simulated annealing algorithm (assisted by numerical analysis in Mathematica) to search the solution space for the global minimum of $C_N$. This allows for the determination of optimal directions for $N$ up to 20.
3. Analytical and Numerical Results: For $N \leq 6$, the authors derive analytical expressions for the measurement directions and bounds. For $N > 7$, they provide numerical results.
4. Infinite-Setting Limit ($N \to \infty$): To resolve the critical value exactly, the authors analyze the limit as $N \to \infty$. They model the measurement directions as points distributed evenly over the Northern Hemisphere of the unit sphere. By dividing the hemisphere into concentric circles and calculating the vector sum of these directions, they derive the asymptotic behavior of the classical bound.

Key Results

• Optimization of Finite Settings: The study confirms that SJWP's inequalities are optimal only for $N=2$ and $N=3$. For $N \geq 4$, the authors find new measurement configurations that yield lower classical bounds $C_N$.
  • For $N=20$, the optimized bound is $C_{20} \approx 0.5073$, deviating from the target $0.5$ by only $0.007$.
• Exact Critical Value: In the limit of infinite settings ($N \to \infty$), the authors analytically demonstrate that the classical bound converges to:
  $$\lim_{N \to \infty} C_N = \frac{1}{2}$$
  Consequently, the quantum violation of this inequality occurs precisely when $V > 1/2$. This successfully recovers the critical value $V_c^S = 1/2$ for the two-qubit Werner state using the inequality approach.
• Comparative Performance: The Supplemental Material indicates that these optimal inequalities possess a stronger capability to detect nonlocality in generalized Werner states, maximal entangled mixed states, and all-versus-nothing states compared to the original SJWP inequalities.

Significance and Claims
The paper claims to resolve a long-standing problem in quantum foundations by providing the first inequality-based derivation of the exact critical value $V_c^S = 1/2$ for the two-qubit Werner state. By establishing the optimal $N$-setting linear EPR-steering inequalities and proving their convergence to $1/2$ in the infinite-setting limit, the work bridges the gap between the theoretical definition of the EPR steering boundary and its demonstration via steering inequalities. The authors note that while the steering boundary is now resolved, the problem of revealing the Bell nonlocality boundary ($V_c^B = 1/KG(3)$) via Bell inequalities remains an open task for future investigation.