matrix factorization notebook, no activations?

[jmschrei]

What is the benefit of adding in another layer if you're using a linear activation? My understanding was that multiple linear layers can be expressed as a single linear layer.

Thanks!

[mli]

fixed, thanks.