Add simple working example of basic TCN / MomentNetwork structure

[JLDC]

Proposed changes

Add a simple example of how a neural network (temporal convolutional network) could be used to construct the moments of a given underlying model.

Types of changes

• [ ] Bugfix (non-breaking change which fixes an issue)
• [x] New feature (non-breaking change which adds functionality)
• [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)

Checklist

• [x] I have read the CONTRIBUTING doc
• [x] I am making a pull request against the main branch (left side). Also you should start your branch off our main.
• [ ] Lint and unit tests pass locally with my changes
• [ ] I have added tests that prove my fix is effective or that my feature works

Further comments

This is really a minimal working example somewhat outside of the package structure. It should be integrated in the package structure and the PR is being created as discussed with @AldoGl. He will make changes to this PR before integrating it.

[AldoGl]

Thank you Jonathan, great work! We'll try and integrate your algorithm soon :)

[codecov-commenter]

Codecov Report

Merging #58 (35d2742) into main (152aa9b) will not change coverage. The diff coverage is n/a.

:exclamation: Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

Additional details and impacted files

[](https://app.codecov.io/gh/bancaditalia/black-it/pull/58?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None) ```diff @@ Coverage Diff @@ ## main #58 +/- ## ======================================= Coverage 91.36% 91.36% ======================================= Files 42 42 Lines 1737 1737 ======================================= Hits 1587 1587 Misses 150 150 ```

[AldoGl]

Today I worked on the code and did the following:

• I moved modules into appropriate folders (loss_functions, test, example)
• I better integrated the code with existing classes (for instance, now the sampler used in the network is a Black-it sampler, and the parameter ranges are specified via a SearchSpace object)
• I added an example notebook (where i first compute the optimal moment function, and then I use it within a method of moments calibration)
• I fixed typing and linting error as much as possible
• I added @JLDC as an author

There is still some work to do, primarily:

• write the rest of the docstrings
• fix all linting errors
• write a test (I added a dummy test, we can start from there)
• consider what should be done to extend the code to multiple moments while preserving existing abstractions within the method of moments loss function

[AldoGl]

@JLDC, could you give access to your fork also to @marcofavoritobi? (no hurry)

[JLDC]

@JLDC, could you give access to your fork also to @marcofavoritobi? (no hurry)

He should already have access. Any maintainer is allowed to edit this pull request. Let me know if I am mistaken and I need to undertake something else.