Witnessing network steerability of every bipartite entangled state without inputs

1. Problem Statement

The paper addresses a fundamental open problem in quantum information theory regarding the relationship between entanglement and quantum steering.

• The Gap: While all steerable states are entangled, the converse is not true in standard scenarios. There exist entangled states (e.g., certain Werner states) that are "unsteerable," meaning their correlations can be explained by a Local Hidden State (LHS) model.
• The Question: Can the steerability of every entangled state be "activated" or demonstrated in some extended scenario?
• The Context: Previous work showed that using multiple copies of a state (superactivation) could reveal steering. This paper investigates whether quantum networks (specifically the swap-steering scenario) can demonstrate steering for all bipartite entangled states, even those that are unsteerable in standard settings.
• The Constraint: The authors aim to achieve this without inputs (i.e., parties perform fixed measurements rather than choosing from a set of inputs), which is the minimal scenario for observing network nonlocality.

2. Methodology

The authors utilize the swap-steering scenario, a network configuration involving two spatially separated parties (Alice and Bob) and two independent sources ($S_1, S_2$).

• Setup:
  • Alice and Bob each receive two subsystems (one from each source).
  • Alice is the "trusted" party (her measurement devices are well-characterized).
  • Bob is untrusted (his devices are treated as a black box).
  • Both parties perform fixed measurements on their subsystems.
  • The goal is to detect if the correlations violate a Separable Outcome-Independent Hidden State (SOHS) model.
• Theoretical Tools:
  • Entanglement Witnesses: The authors map known entanglement criteria (PPT and CCN) to network steering inequalities.
  • Entanglement Swapping: They leverage the mechanism where Bob's measurement on his two subsystems projects Alice's two subsystems into a specific state.
  • Tomography: For the general case, they allow the trusted party (Alice) to perform a tomographically complete set of measurements to reconstruct the state.

3. Key Contributions and Results

The paper presents three main theoretical results, progressively generalizing the scope of detectable states:

A. Witnessing NPT States (Negative Partial Transpose)

• Target: All bipartite states with a Negative Partial Transpose (NPT).
• Method: The authors construct a linear inequality ($S_{\rho:NPT}$) based on the eigenvector $|\eta\rangle$ corresponding to the negative eigenvalue of the partial transpose of the target state $\rho$.
• Result: They prove that for any NPT state, the value of the witness exceeds the maximum possible value attainable by an SOHS model ($\beta_{SOHS} = 0$).
• Significance: This confirms that all NPT states are network steerable in the swap scenario without inputs.

B. Witnessing CCN-Violating States (including Bound Entanglement)

• Target: A large class of states violating the Computable Cross-Norm (CCN) criterion. This class includes some bound entangled states (states that are entangled but have a positive partial transpose).
• Method: A different linear inequality ($S_{\rho:CCN}$) is constructed based on the Schmidt decomposition coefficients of the state in the operator space.
• Result: The quantum value of this witness exceeds the SOHS bound ($\beta_{SOHS} = 1/d$).
• Key Finding (Unbounded Gap): The authors demonstrate an unbounded gap between the quantum value and the classical (SOHS) bound.
  • The ratio of the quantum value to the classical bound scales linearly with the dimension $d$ ($Ratio \propto d$).
  • This is the first instance of an unbounded gap between network local and nonlocal correlations in a scenario with no inputs.
  • Implication: This large gap provides significant robustness against experimental noise and imperfections, making high-dimensional swap-steering easier to observe experimentally.

C. Witnessing Every Bipartite Entangled State

• Target: Arbitrary bipartite entangled states (including those that might not violate PPT or CCN in specific bases, though the paper argues all entangled states have a witness).
• Method:
  • Utilizing the Hahn-Banach separation theorem, the authors note that for any entangled state $\tilde{\rho}$, there exists an entanglement witness $W_{\tilde{\rho}}$ such that $\text{Tr}(W_{\tilde{\rho}}\tilde{\rho}) < 0$.
  • They generalize the swap-steering scenario to allow the trusted party (Alice) to perform tomography (a full set of measurements) on her subsystems.
  • They construct a linear inequality ($S_{\tilde{\rho}}$) where the coefficients are directly derived from the entanglement witness $W_{\tilde{\rho}}$.
• Result: They prove that $S_{\tilde{\rho}} > 0$ for any entangled state, while $S_{\tilde{\rho}} \leq 0$ for any SOHS model.
• Conclusion: Every bipartite entangled state is network steerable in the swap-steering scenario, provided the trusted party can perform tomography. This resolves the long-standing question of whether every entangled state can demonstrate steering in some network configuration.

4. Significance and Impact

1. Resolution of an Open Problem: The work definitively answers that the steerability of every bipartite entangled state can be activated in a quantum network. There is no "unsteerable" entangled state in the context of swap-steering.
2. Unbounded Gap Discovery: The discovery of an unbounded gap between local and nonlocal correlations without inputs is a major foundational breakthrough. It implies that quantum strategies can scalably outperform classical strategies in operational tasks without requiring complex input choices.
3. Experimental Feasibility:
   • The proposed witnesses are linear and easy to implement.
   • The setup requires only linear optics (for Bell measurements and tomography).
   • The unbounded gap suggests that even in noisy, high-dimensional systems, the violation of local models remains detectable, reducing the precision and data requirements for experiments.
4. One-to-One Correspondence: The paper establishes a phenomenological link: in the swap-steering scenario, the ability to detect entanglement in the trusted party's subsystem is equivalent to the network being steerable. If the state is entangled, it is steerable; if separable, it is not.

5. Conclusion

Shubhayan Sarkar demonstrates that by moving from standard steering scenarios to networked swap-steering scenarios, the limitations of standard quantum steering are removed. By constructing specific linear witnesses for NPT states, CCN-violating states, and general entangled states (via tomography), the paper proves that entanglement is sufficient for network steerability. The identification of an unbounded gap in a no-input scenario further highlights the robustness and scalability of quantum correlations in networks, offering a promising pathway for experimental verification and quantum information applications.