No multiplicaton with inputs in integrated gradient?

sicara / tf-explain

Interpretability Methods for tf.keras models with Tensorflow 2.x

https://tf-explain.readthedocs.io

MIT License

1.02k stars 110 forks source link

No multiplicaton with inputs in integrated gradient? #128

Open Holt59 opened 4 years ago

Holt59 commented 4 years ago

In the integrated gradient paper, the formula for the ith dimension is:

But in your implementation, it looks like you never multiply with images (since you assume a zero-baseline). Is that intended? If so, could you explain the reasoning behind this?