This tutorial introduces an extra tool for the Teacher-student scheme to systematically compare different QNN architectures and to evaluate their relative expressive power. Along with the averaged operator size (arXiv:2011.07698) and the Fourier coefficients (arXiv:2105.01477), this tutorial introduces the tools to explore specifically the effect of the processing and measurement operators on the expressive power of quantum models.
It is based on the paper "The effect of the processing and measurement operators on the expressive power of quantum models" by Katerina Gratsea and Patrick Huembeli. For more details on the Teacher-student scheme have a look in our work by Teacher-student scheme and the part-1: Teacher-student scheme of the current tutorial.
All code is written in Python. Libraries required with the versions used: pennylane 0.18.0, numpy 1.21.4, matplotlib 3.5.1.