Review of methods for time-series explainability of Neural Networks.
The demo presented here uses pretrained models for speech-based age group classification.
In order to install, run the following chain of commands: 0. First, get the submodules
git submodule update --init --recursive
- Create python virtualenv and install the requirements:
python3 -m virtualenv venv
source venv/bin/activate
pip install -r requirements.txt
- Inside the virtualnevironment install the ftdnn submodule:
./install.sh
For all the details about the algorithms used please check the demo notebook time_series_xplain.ipynb
.
There is little point in copy-pasting its contents here.