Adding dynamic reference functionality to tensorflow DeepExplainer #322
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Added functionality to determine the reference on-the-fly according to the input data. This is implemented by allowing the user to supply a function as the
data
parameter. I benchmarked the change using the genomics model+simulated data from the DeepLIFT repo, and the results look good. The reference generated for each sequence is a collection of dinucleotide-shuffled versions of the sequence. A Colab notebook comparing the output using the dynamic reference API and the standard API (where the standard API requires instantiating a new DeepExplainer object for each new reference) is here: https://gist.github.com/AvantiShri/8a3a0a03f4c4a578ee7909e3989467cc.I added an example notebook to the
notebooks
folder, and bumped the version to reflect the change - feel free to roll back these modifications at your discretion.