How exactly does tensor contraction work with higher rank tensors?

83 Views Asked by At

I'm an undergraduate student doing my best to learn about tensor networks, but I'm stuck on Equation (2) of this paper "A Practical Introduction to Tensor Networks: Matrix Product States and Projected Entangled Pair States" (https://arxiv.org/abs/1306.2164), which is

$$ F_{\gamma\omega\rho\sigma} = \sum^D_{\alpha,\beta,\delta,\nu,\mu=1} A_{\alpha\beta\delta\sigma} B_{\beta\gamma\mu} C_{\delta\nu\mu\omega}E_{\nu\rho\alpha} $$

Are the variables that the author chose to contract over arbitrary? Or are they a result of coinciding dimensions, like how $\beta$ is the second dimension of A and first dimension of B, or how the same relationship is present for $\nu$ in C and E?

It may be that I'm focusing way too hard on the details, but I want to know how the indices of C and E collapse/contract to be compatible with B and so on (IF it is the case that this contraction happens because indices coincide).

Thank you!