Prerequisites & Notation

Before You Begin

This chapter requires familiarity with PyTorch nn.Module (Chapter 26), CNNs (Chapter 27), and complex number representation in PyTorch tensors (Chapter 12). Knowledge of wireless channel models is beneficial.

PyTorch nn.Module and training loop (Chapter 26)(Review ch26)
Self-check: Can you write a custom nn.Module and train it?
CNNs: Conv2d, BatchNorm, residual blocks (Chapter 27)(Review ch27)
Self-check: Can you build a U-Net or ResNet?
PyTorch complex tensors and autograd (Chapter 12)(Review ch12)
Self-check: Can you create a complex tensor and compute gradients through complex operations?

Notation for This Chapter

Symbol	Meaning	Introduced
$\\mathbf{z} \\in \\mathbb{C}^n$	Complex-valued vector	s01
$\\Re(\\cdot), \\Im(\\cdot)$	Real and imaginary parts	s01
$\\mathcal{A}(\\cdot)$	Forward model (physics operator)	s03
$\\partial / \\partial z^*$	Wirtinger derivative (conjugate gradient)	s01

← Ch 27 Complex-Valued Networks