WIP: Fine Grained Workflow API Tutorial

[psfoley]

This PR:

• Represents a first in a tutorial series on demystifying federated learning - which will walk through the considerations when implementing an FL framework at a low level
• Demonstrates through the OpenFL Workflow API the importance of:
  • Security for FL
  • Reducing Communication cost
• Introduces methods for manipulating model weights and applying compression
• Demonstrates ease of implementing exfiltration attacks through models
• Performs a comparison between compression and it's effect on final model accuracy

The tutorial can be launched through Google Colab here:

Todo:

• Refactor low level interfaces from notebook into openfl.experimental namespace (i.e. openfl.experimental.model, openfl.experimental.compression)