Adds a novel operator. The fusion operator modifies the inputs $u$ by using all dimensions of all sensors as input to a linear layer. The architecture utilizes a deep residual network to process the stacked $y$ and preprocessed $u$.
How does it solve the problem?
Implements Fusion Operator.
How are the changes tested?
Added unit tests for shapes and integration test (convergence study).
Checklist for Contributors
[x] Scope: This PR tackles exactly one problem.
[x] Conventions: The branch follows the feature/title-slug convention.
[x] Conventions: The PR title follows the Bugfix: Title convention.
[x] Coding style: The code passes all pre-commit hooks.
[x] Documentation: All changes are well-documented.
[x] Tests: New features are tested and all tests pass successfully.
[x] Changelog: Updated CHANGELOG.md for new features or breaking changes.
[x] Review: A suitable reviewer has been assigned.
Checklist for Reviewers:
[ ] The PR solves the issue it claims to solve and only this one.
[ ] Changes are tested sufficiently and all tests pass.
[ ] Documentation is complete and well-written.
[ ] Changelog has been updated, if necessary.
Open Questions
Should the locations $x$ of the sensors also be introduced in this stack?
For the integration test this reduces the required number of epochs from 896 to 573.
Feature: Fusion Operator
Description
Adds a novel operator. The fusion operator modifies the inputs $u$ by using all dimensions of all sensors as input to a linear layer. The architecture utilizes a deep residual network to process the stacked $y$ and preprocessed $u$.
How does it solve the problem?
How are the changes tested?
Checklist for Contributors
feature/title-slug
convention.Bugfix: Title
convention.Checklist for Reviewers:
Open Questions