Make ANNs compatible with Experiment

[tqtg]

Description

We now can compare recommendation performance of an ANN with other models under the same Experiment as demonstrated in the example. Alternatively, ANN can be built from a trained model as before. They're now flexible for different scenarios.

Related Issues

Checklist:

• I have added tests.
• I have updated the documentation accordingly.
• I have updated README.md (if you are adding a new model).
• I have updated examples/README.md (if you are adding a new example).
• I have updated datasets/README.md (if you are adding a new dataset).

[lthoang]

@tqtg we didn't check if a model is fitted before. Is it better if we just show warnings instead of throwing exception when a model is_fitted? This will not affect execution if someone reuses their model.