Now we are finally ready to implement the matrix factorization model in TensorFlow. Let us do this and see if we can improve the baseline by implicit. Implementing ALS in TensorFlow is not an easy task: it is better suited for gradient-based methods such as SGD. This is why we will do exactly that, and leave ALS to specialized implementations.
Here we implement the formula from the previous sections:
.Recall that the objective there was the following:
Note that in this objective we still have the squared error, ...
