Remove or mitigate the memoryless-device assumption

Determine whether the memoryless-device assumption can be completely removed, or otherwise mitigated, in loophole-free tests of non-contextuality inequalities, while accounting for the additional resources required by such a modification.

Background

The proposed semi-device-independent randomness-expansion protocol assumes that the preparation and measurement devices are memoryless: each round is independent of previous rounds, and outcomes do not depend on prior settings or outcomes. The paper identifies this assumption as necessary for its loophole-free contextuality test and as a drawback shared by cryptographic schemes based on quantum contextuality.

The authors note that recent work suggests the assumption might be removable, but potentially at the expense of additional resources. Establishing whether memory effects can be eliminated or controlled without undermining loophole-free contextuality tests and secure randomness expansion is therefore left as a direction for future research.