explainable-continual-learning
eXplaning continual learning with SHAP
Drifting explanations in continual learning
To reproduce the experiments we provided the scripts in the file scripts.sh.
The scripts trains all the models and saves the results in the folder experiment_folder provided by command line.
The script.sh file does not include the path to the experiment folder, you can add it pointing to the folder of your choice.
Also, for the SSC experiments you need to provide the path containing the preprocessed dataset, that you can download here.
You can select the explanator you want by using either shap or lift in the command line arguments.
To get all the plots of the paper, you should run explanation_over_time.py by providing the same experiment folder as command line input.
For each benchmark, continual learning strategy, model and explanator, the script will create different plots for the different metrics.