Normand Beaudry

Removing assumptions about devices and signals from quantum cryptography protocols is an important task. Often it is assumed that measurements are memoryless and each signal is measured independently. We aim to remove this assumption by considering a related problem involving what we call quantum exams: First assume a student has a quantum memory which has an encoding of a classical string. Consider that the student is asked to produce a small subset of the string (i.e. and exam), and then is asked for another subset of the string. If the student can make a good estimate in the first exam, he should be able to also make a good estimate in the second exam. We examine this problem in the case of a classical or quantum memory for the student, and its relation to quantum cryptography.