All Bell inequalities with a quantum advantage in the minimal scenario are self-testing

In the simplest Bell scenario, every linear Bell inequality that shows a quantum advantage uniquely identifies (self‑tests) the underlying quantum state and measurements.

All Bell inequalities with a quantum advantage in the minimal scenario are self-testing

Introduction

The minimal Bell scenario—two parties, two measurement settings per party, and binary outcomes—has long served as a testbed for foundational questions in quantum non‑locality. While the set of extremal quantum correlations in this setting is completely characterized, the family of Tsirelson (quantum) inequalities that bound this set remains only partially understood. A recent pre‑print by Liu, Zingarofalo, Barizien, and Bancal tackles this gap by showing that every Bell inequality that yields a quantum advantage in the minimal scenario is self‑testing. In other words, the inequality’s maximal quantum violation can be achieved by a single, extremal correlation, which in turn uniquely determines the underlying quantum realization.

From Tsirelson Inequalities to Self‑Testing

A Tsirelson inequality is a linear constraint on the observed correlation vector that is satisfied by all quantum correlations but can be violated by supra‑quantum (no‑signalling) ones. The authors first examine the mapping between a given inequality and the set of quantum correlations that maximize it. By exploiting the known geometry of the quantum set in the minimal scenario, they prove that any inequality with a quantum advantage admits exactly one extremal quantum point that attains the maximal value. This point is an extremal point of the quantum set, meaning it cannot be expressed as a convex combination of other quantum correlations.

Because extremal points correspond to pure, projective measurements on a pure bipartite state, the uniqueness of the maximizer implies a self‑testing statement: observing the maximal violation of the inequality certifies, up to local isometries, the specific state and measurement operators used. The result therefore provides a complete classification of linear self‑tests in the minimal scenario: they are precisely the Tsirelson inequalities with quantum advantage.

Limited Families of Non‑Unique Maximizers

While most inequalities are uniquely maximized, the authors identify a narrow exception. They show that at most a two‑dimensional family of inequalities can share the same non‑local extremal correlation as a maximizer. These families correspond to polygonal facets of the quantum set generated by qubit states with arbitrary entanglement. Geometrically, the set of correlations generated by a fixed entangled qubit pair traces out a curved surface; the intersecting linear facets form polygons whose edges are supported by the same extremal point.

This observation leads to an important structural insight: the quantum set in the minimal scenario is not self‑dual. Self‑duality would require that the set of supporting hyperplanes (the inequalities) be in one‑to‑one correspondence with the extremal points, which is violated by the identified polygonal families.

Practical Implications

The self‑testing property of all quantum‑advantage Bell inequalities simplifies device‑independent protocols in the minimal scenario. Any experiment that achieves the maximal quantum value of a Bell inequality automatically certifies the underlying resources without additional assumptions. Moreover, the characterization of the exceptional polygonal families informs the design of robust tests: by steering away from those specific facets, one can ensure unique self‑testing.

Conclusion

Liu and collaborators provide a clear and comprehensive answer to a long‑standing question: every linear Bell inequality that demonstrates a quantum advantage in the simplest Bell scenario is a self‑test. Their work not only clarifies the geometry of the quantum set—highlighting its non‑self‑dual nature—but also strengthens the practical toolbox for device‑independent quantum information tasks. Future research may extend these techniques to scenarios with more inputs, outputs, or parties, probing whether the tight link between quantum advantage and self‑testing persists in higher‑dimensional settings.

Source

Read original source

Why we picked this

Accepted from curated QuantumUnzipped source (relevance scoring disabled).

← Back to all articles