Hardy-type self-testing and exposedness of tripartite GHZ correlations

This paper demonstrates that, contrary to the bipartite case, the tripartite GHZ correlation maximizing Hardy's paradox success probability is an exposed point of the quantum set that simultaneously self-tests the GHZ state and coincides with the maximal violation of the Mermin inequality, thereby unifying logical-paradox and Bell-inequality routes to nonlocality in the multipartite setting.

The Gist

Imagine you are trying to prove that a group of friends is secretly communicating, even though they are in separate rooms and cannot talk to each other. In the world of quantum physics, this is called nonlocality. Scientists usually prove this in two different ways:

1. The "Math Test" (Bell Inequalities): You give them a complex math problem. If they are just guessing or using hidden notes (classical physics), they will fail. If they are using "spooky" quantum magic, they will get a score higher than the math allows.
2. The "Logic Puzzle" (Hardy's Paradox): Instead of a score, you look for a specific pattern of answers. You say, "If you get these three answers, you must get this fourth answer. But if you get the fourth answer, it's impossible for you to be using hidden notes." It's a logical trap that only quantum mechanics can spring.

For a long time, scientists thought these two methods were very different, especially when looking at two people (the "bipartite" scenario). They found that the "Math Test" winners were like sharp, exposed peaks on a mountain range—you could easily point to them with a straight line. But the "Logic Puzzle" winners were like hidden valleys or flat plateaus; they were special, but you couldn't point to them with a single straight line. They were "non-exposed."

The Big Discovery
This paper asks: "Does this difference still exist if we have three people instead of two?" (The "tripartite" scenario).

The authors, Smritikana Patra, Soumyajit Pal, and Ranendu Adhikary, say: No, the rules change completely.

Here is what they found, using simple analogies:

1. The Three-Person Logic Trap

They set up a three-person version of the "Logic Puzzle" (Hardy's paradox). They asked: "What is the best possible quantum strategy to win this puzzle?"

• The Result: The best strategy turns out to be a very famous quantum state called the GHZ state (Greenberger–Horne–Zeilinger). Think of this as three coins that are magically linked so that if you flip them, they always land in a specific, synchronized pattern.
• The Proof: They proved that if you see this specific winning pattern, you know for a fact that the three people must be sharing this GHZ state and using specific measurement tools. This is called self-testing. It's like seeing a unique fingerprint and knowing exactly which person made it, without ever seeing the person.

2. The Mountain Peak Surprise

Here is the most surprising part. In the two-person world, the "Logic Puzzle" winners were hidden valleys (non-exposed). But in the three-person world, the authors proved that the winner of the Logic Puzzle is actually a sharp, exposed peak.

• The Analogy: Imagine a mountain range. In the two-person world, the Logic Puzzle winner was a flat spot on a cliff that you couldn't touch with a ruler. In the three-person world, the Logic Puzzle winner is a sharp, jagged peak. You can place a flat board (a "supporting hyperplane") right under it, and it touches only that one point.
• Why it matters: This means the "Logic Puzzle" and the "Math Test" are actually pointing to the exact same spot. The correlation that wins the Logic Puzzle is also the one that breaks the "Math Test" (the Mermin inequality) the most.

3. The "Real World" Check

In real life, experiments aren't perfect. There is always a little bit of noise or error. You can't get a perfect "zero" probability in a lab.

• The authors checked if their "Logic Puzzle" proof still works if the answers are slightly messy.
• The Result: Yes! Even if the experiment is slightly imperfect (within a very small margin of error), the proof still holds. If the results are close enough to the ideal pattern, you can still be confident that the three people are sharing the GHZ state.

Summary

In the world of two people, "Logic Puzzles" and "Math Tests" for quantum weirdness lead to different geometric shapes (one hidden, one exposed).

In the world of three people, the authors discovered that these two paths merge. The "Logic Puzzle" winner is no longer a hidden valley; it is a sharp, exposed peak that is identical to the "Math Test" winner. They both certify the same magical three-person connection (the GHZ state).

This changes our understanding of the geometry of quantum reality, showing that adding just one more person to the mix fundamentally changes how these quantum secrets are revealed.

Technical Summary

Technical Summary: Hardy-type Self-Testing and Exposedness of Tripartite GHZ Correlations

Problem Statement
The paper addresses the geometric characterization of quantum nonlocality, specifically comparing two distinct routes to witnessing nonlocality: Bell-inequality violations and logical contradictions (Hardy's paradox). In the bipartite two-input two-output scenario $((2, 2, 2))$, a known geometric distinction exists: correlations achieving maximal CHSH violation are "exposed" points of the quantum set (uniquely selected by a supporting hyperplane), whereas correlations achieving maximal Hardy success probability are "non-exposed" extremal points. The authors investigate whether this bipartite intuition holds in the tripartite $((3, 2, 2))$ scenario, where the Greenberger–Horne–Zeilinger (GHZ) state is central. Specifically, they ask if the tripartite Hardy-maximal correlation is also non-exposed or if it exhibits different geometric properties.

Methodology
The authors employ a combination of analytical proofs and numerical optimization techniques:

1. Self-Testing Framework: They utilize the standard definition of self-testing, where an observed correlation uniquely determines the underlying quantum state and measurements up to local isometries. They first establish that the maximum Hardy success probability $p^H_3$ for a tripartite system is $1/8$, achievable only by the GHZ state with specific observables. They prove that this bound holds for arbitrary finite dimensions by decomposing general states into qubit subspaces.
2. Geometric Analysis (Exposedness): To determine if the Hardy-maximal correlation is an exposed point, the authors formulate a linear programming problem. They search for a Bell functional (linear functional) that is uniquely maximized by the Hardy correlation $\vec{P}_H$. This involves constructing a Bell operator $W$ and verifying that $\vec{P}_H$ is the unique optimizer among all quantum correlations.
3. Connection to Mermin Inequality: They analytically compare the Bell functional derived from the Hardy optimization with the Mermin inequality operator to check for equivalence.
4. Robustness Analysis: Recognizing that experimental data cannot satisfy strict zero-probability constraints, they apply the SWAP method combined with the NPA (Navascués-Pironio-Acín) hierarchy. They formulate Semidefinite Programming (SDP) problems to compute the worst-case fidelity between an unknown state and the ideal GHZ state, given small deviations ($\epsilon_1, \epsilon_2$) from the ideal Hardy constraints.

Key Contributions and Results

• Tripartite Hardy Self-Testing: The authors prove that the quantum correlation attaining the maximal Hardy success probability ($p^H_3 = 1/8$) in the $((3, 2, 2))$ scenario self-tests the tripartite GHZ state and the associated local measurements. This provides a device-independent certification of the GHZ realization based on a logical paradox rather than a Bell inequality.
• Exposedness of the Hardy Point: Contrary to the bipartite case, the paper demonstrates that the tripartite Hardy-maximal correlation is an exposed extremal point of the quantum correlation set. They explicitly construct a Bell functional (a linear combination of correlators) that is uniquely maximized by this correlation.
• Hardy–Mermin Coincidence: The study reveals that the Bell functional maximizing the Hardy correlation is equivalent (up to relabeling) to the Mermin inequality. Consequently, the correlation that maximizes the logical Hardy paradox is identical to the one that maximally violates the Mermin inequality. This unifies the logical-paradox and Bell-inequality routes to nonlocality in the tripartite GHZ scenario, showing they select the same exposed boundary point.
• Robust Self-Testing: The authors establish a robust version of the self-test. Using the NPA hierarchy (level $\ell=6$), they show that if the Hardy zero-probability constraints are satisfied with deviations up to $\epsilon_2 \approx 10^{-4}$ and the success probability is close to optimal, the underlying state and measurements remain quantitatively close to the ideal GHZ realization. Specifically, the fidelity remains above $0.5$ for small deviations in the success probability ($\epsilon_1 \lesssim 0.005$).

Significance
The paper claims that these results reveal a qualitative geometric difference between bipartite and tripartite Hardy-type nonlocality. While bipartite Hardy correlations on the no-signaling boundary are known to be non-exposed, the tripartite case demonstrates that Hardy-type logical nonlocality can lead to exposed quantum boundary points. This challenges the generality of the bipartite intuition regarding the geometric nature of logical paradoxes.

Furthermore, the work suggests that in the tripartite GHZ scenario, the distinction between "paradox-based" and "inequality-based" nonlocality is less sharp geometrically, as both converge to the same exposed point. The authors conclude that this finding motivates further investigation into the exposedness of maximal Hardy correlations in general $((N, 2, 2))$ scenarios and whether similar coincidences exist for higher-party Mermin-Ardehali-Belinskii-Klyshko inequalities. The robustness analysis places Hardy-type self-testing on an operational footing comparable to Bell-inequality-based methods, making it viable for device-independent protocols where ideal constraints cannot be perfectly met.