add scala notebooks & starting scripts.

[Calculuser]

No description provided.

[megaSpoon]

the same modifications are already applied to these notebooks as those in python language. Should I add the changes in this PR or just start a new PR to do it?

Also overall, no big errors to be fixed other than some dependency missing and dataset path not found exceptions. I will put more details to describe it later in the commits.

[megaSpoon]

The loss curve of autoencoder notebook tutorial is not correct. I'm finding why this malfunction happens.

[Calculuser]

Thanks for your update. @megaSpoon

In fact, I will start new PRs to finish Scala tutorials later according to your modifications in python version. So, I think you do not need to add changes in this PR.

If you find the reason why the loss curve is not correct, please contact me. Thanks.