Notebooks/Tutorial_03_implement_dynamask.ipynb

{"cells":[{"cell_type":"markdown","metadata":{},"source":["# Tutorial 3 - Dynamask\n","\n","In this tutorial we we create a Dynamask explainer object and use it to explain a test record. The explainer is then saved to disk and can be given to someone else to view in the [Interpretability Suite App](https://vanderschaarlab-demo-interpretabi-interpretability-suite-1uteyn.streamlit.app/).\n","\n","We will be explaining the predictions of pytorch convolutional neural net that we have trained and saved separately on an engine noise dataset from IEEE World Congress on Computational Intelligence, 2008. The Interpretability.models module provides a a pytorch model for this that is compatible with trained models `state_dict`s available on the Google Drive link below.\n","\n","### Import the relevant modules"]},{"cell_type":"code","execution_count":1,"metadata":{},"outputs":[],"source":["# Standard\n","import os\n","import pathlib\n","\n","# Third Party\n","import numpy as np\n","import torch\n","\n","# Intperpretability\n","from interpretability.models.recurrent_neural_net import ConvNet\n","from interpretability.interpretability_models import dynamask_explainer\n","from interpretability.interpretability_models.utils import io"]},{"cell_type":"markdown","metadata":{},"source":["### Load the data \n","Load the data and split it into the train and the test examples we will explain. This cell will download the data from the `root_url` and save it to a subdirectory in the folder this notebook is being run."]},{"cell_type":"code","execution_count":2,"metadata":{},"outputs":[],"source":["# LOADS\n","def load_forda_data():\n","    def readucr(filename):\n","        data = np.loadtxt(filename, delimiter=\"\\t\")\n","        y = data[:, 0]\n","        x = data[:, 1:]\n","        return x, y.astype(int)\n","\n","    root_url = \"https://raw.githubusercontent.com/hfawaz/cd-diagram/master/FordA/\"\n","\n","    x_train, y_train = readucr(root_url + \"FordA_TRAIN.tsv\")\n","    x_test, y_test = readucr(root_url + \"FordA_TEST.tsv\")\n","\n","    x_train = x_train.reshape((x_train.shape[0], x_train.shape[1], 1))\n","    x_test = x_test.reshape((x_test.shape[0], x_test.shape[1], 1))\n","\n","    idx = np.random.permutation(len(x_train))\n","    x_train = x_train[idx]\n","    y_train = y_train[idx]\n","\n","    y_train[y_train == -1] = 0\n","    y_test[y_test == -1] = 0\n","\n","    return x_train, y_train, x_test, y_test\n","\n","\n","X_train, y_train, X_explain, y_explain = load_forda_data()"]},{"cell_type":"markdown","metadata":{},"source":["### Download the trained model from Google Drive\n","\n","You could train your own model using the ConvNet class and load it here, but we have trained one already.\n","\n","Download the model using this link: https://drive.google.com/file/d/173vniHegUSGmdC6fKCLupynRoxEdz9Ko/view?usp=sharing and save it in a location matching the path `TRAINED_MODEL_STATE_PATH` below (This is the same model as in \"Tutorial_02_implement_simplex_time_series\", so if you have already done so, there is no reason to download it again). The default location is the `\"resources/saved_models\"` folder inside the root Interpretability directory.\n","\n","\n","### Load the model"]},{"cell_type":"code","execution_count":3,"metadata":{},"outputs":[],"source":["## Load the model\n","model = ConvNet()\n","\n","def load_trained_model(model, trained_model_state_path, device='cpu'):\n","    model.load_state_dict(torch.load(trained_model_state_path, map_location=torch.device(device)))\n","    model.eval()\n","    return model\n","\n","DEVICE = \"cpu\"\n","\n","root_path = pathlib.Path.cwd().parents[0]\n","saved_models_path = root_path / \"resources/saved_models\"\n","TRAINED_MODEL_STATE_PATH = saved_models_path / \"model_cv1_2.pth\"\n","model = load_trained_model(model, TRAINED_MODEL_STATE_PATH, device=DEVICE)"]},{"cell_type":"markdown","metadata":{},"source":["### Initialize Dynamask\n","Initialize the explainer object by passing the predictive model, the perturbation method to use, and whether or not to use a `Mask` or `GroupMask`."]},{"cell_type":"code","execution_count":4,"metadata":{},"outputs":[],"source":["available_perturbation_method = [\n","    \"fade_moving_average\",\n","    \"gaussian_blur\",\n","    \"fade_moving_average_window\",\n","    \"fade_moving_average_past_window\",\n","    \"fade_reference\",\n","]\n","\n","my_explainer = dynamask_explainer.DynamaskExplainer(\n","    model, available_perturbation_method[1], group=False\n",")"]},{"cell_type":"markdown","metadata":{},"source":["### Fit the explainer\n","\n","Fit the explainer on the training data. This makes explanations of the test data available in the subsequent step."]},{"cell_type":"code","execution_count":5,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["Fitting Dynamask\n"]},{"name":"stderr","output_type":"stream","text":["/home/rob/.local/lib/python3.10/site-packages/torch/nn/modules/conv.py:306: UserWarning: Applied workaround for CuDNN issue, install nvrtc.so (Triggered internally at ../aten/src/ATen/native/cudnn/Conv_v8.cpp:80.)\n","  return F.conv1d(input, weight, bias, self.stride,\n"]},{"name":"stdout","output_type":"stream","text":["====================================================================================================\n","The optimization finished: error = 3.17e-05 ; size regulator = 1.49e-07 ; time regulator = 0.00224 ; time elapsed = 3.72 s\n","====================================================================================================\n","\n"]}],"source":["available_loss_functions = [\n","    \"cross_entropy\",\n","    \"log_loss\",\n","    \"log_loss_target\",\n","    \"mse\",\n","]\n","\n","explain_id = 1\n","\n","my_explainer.fit(\n","    explain_id,\n","    X_train,\n","    loss_function=available_loss_functions[0],\n","    target=y_train[explain_id],\n","    area_list=np.arange(0.1, 0.5, 0.1),\n",")\n","feature_num, time_step_num = X_train[0].shape"]},{"cell_type":"markdown","metadata":{},"source":["### Get the explanation\n","Get the explanation for the record selected above. We can also choose whether to smooth the mask or not."]},{"cell_type":"code","execution_count":6,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["   0         1    2    3        4    5    6        7    8    9    ...  490  \\\n","0  0.0  0.002335  0.0  0.0  0.00217  0.0  0.0  0.00218  0.0  0.0  ...  1.0   \n","\n","   491  492  493  494  495  496  497  498  499  \n","0  1.0  1.0  1.0  1.0  1.0  1.0  1.0  1.0  1.0  \n","\n","[1 rows x 500 columns]\n"]}],"source":["print(\n","    my_explainer.explain(\n","        # ids_feature=[i for i in range(5)],\n","        # ids_time=[i for i in range(5)],\n","        smooth=False,\n","        get_mask_from_group_method=\"extremal\",\n","        extremal_mask_threshold=0.01,\n","    ).feature_importances\n",")"]},{"cell_type":"markdown","metadata":{},"source":["### Plot the explanation\n","Plot the feature importance over time. This produces an image as a png."]},{"cell_type":"code","execution_count":7,"metadata":{},"outputs":[{"data":{"image/png":"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","text/plain":["<Figure size 640x480 with 2 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["my_explainer.summary_plot()"]},{"cell_type":"markdown","metadata":{},"source":["### Save the explainer to file\n","This file can now be uploaded to the [Interpretability Suite App](https://vanderschaarlab-demo-interpretabi-interpretability-suite-1uteyn.streamlit.app/). This provides a non-programtic interface with which to view the various explanations, allowing you to send the explainer to a colleague who is less fluent in python."]},{"cell_type":"code","execution_count":8,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["Saving explainer to: /home/rob/Documents/projects/Interpretability/Notebooks/my_new_forda_conv_dynamask_explainer.p\n"]}],"source":["io.save_explainer(\n","    my_explainer, os.path.abspath(\"my_new_forda_conv_dynamask_explainer.p\")\n",")"]}],"metadata":{"kernelspec":{"display_name":"Python 3.8.13 ('interp')","language":"python","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.10.14"},"orig_nbformat":4,"vscode":{"interpreter":{"hash":"6fd73f071793638ac14baf0ff0f19e5ab81431475f40d47f0df0002312a62017"}}},"nbformat":4,"nbformat_minor":2}