issue #724 improved link to RISE paper- the conference publicaiton, n…

…ot the argxiv one
dianna-ai · Apr 17, 2024 · 56be516 · 56be516
1 parent 7d9e813
commit 56be516
Show file tree

Hide file tree

Showing 5 changed files with 10 additions and 37 deletions.
diff --git a/tutorials/explainers/RISE/rise_imagenet.ipynb b/tutorials/explainers/RISE/rise_imagenet.ipynb
@@ -10,9 +10,8 @@
     "\n",
     "This notebook demonstrates how to apply the RISE explainability method on pretrained ImageNet model using a bee image. It visualizes the relevance scores for all pixels/super-pixels by displaying them on the image.<br>\n",
     "\n",
-    "RISE is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates importance empirically by probing the model with randomly masked versions of the input image and obtaining the corresponding outputs.<br>\n",
-    "\n",
-    "More details about this method can be found in the paper https://arxiv.org/abs/1806.07421.<br>\n"
+    "[RISE](http://bmvc2018.org/contents/papers/1064.pdf) is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates importance empirically by probing the model with randomly masked versions of the input image and obtaining the corresponding outputs.<br>\n",
+    "\n"
    ]
   },
   {

diff --git a/tutorials/explainers/RISE/rise_mnist.ipynb b/tutorials/explainers/RISE/rise_mnist.ipynb
@@ -2,7 +2,6 @@
  "cells": [
   {
    "cell_type": "markdown",
-   "id": "92d7325c",
    "metadata": {},
    "source": [
     "<img width=\"150\" alt=\"Logo_ER10\" src=\"https://user-images.githubusercontent.com/3244249/151994514-b584b984-a148-4ade-80ee-0f88b0aefa45.png\">\n",
@@ -11,14 +10,11 @@
     "\n",
     "This notebook demonstrates how to apply the RISE explainability method on a pretrained binary MNIST model using a hand-written digit image. It visualizes the relevance attributions for each pixel/super-pixel by displaying them on top of the input image.<br>\n",
     "\n",
-    "RISE is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates the relevance empirically by probing the model with randomly masked versions of the input image to obtain the corresponding outputs.<br>\n",
-    "\n",
-    "More details about this method can be found in the paper https://arxiv.org/abs/1806.07421.<br>"
+    "[RISE](http://bmvc2018.org/contents/papers/1064.pdf) is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates the relevance empirically by probing the model with randomly masked versions of the input image to obtain the corresponding outputs.<br>"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "293e83ff",
    "metadata": {},
    "source": [
     "### Colab Setup"
@@ -27,7 +23,6 @@
   {
    "cell_type": "code",
    "execution_count": 1,
-   "id": "8d3eb1d8",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -46,7 +41,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "fc18a3f6",
    "metadata": {},
    "source": [
     "### Libraries"
@@ -55,7 +49,6 @@
   {
    "cell_type": "code",
    "execution_count": 1,
-   "id": "84d588b8",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -73,7 +66,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "c15254cf",
    "metadata": {},
    "source": [
     "#### 1 - Loading the model and the dataset\n",
@@ -82,7 +74,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "f7ee0349",
    "metadata": {},
    "source": [
     "Load saved binary MNIST data."
@@ -91,7 +82,6 @@
   {
    "cell_type": "code",
    "execution_count": 2,
-   "id": "dde4c313",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -104,7 +94,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "4de48ca0",
    "metadata": {},
    "source": [
     "Load the pretrained binary MNIST model and define a model runner."
@@ -113,7 +102,6 @@
   {
    "cell_type": "code",
    "execution_count": 3,
-   "id": "7f3d743e",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -134,7 +122,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "1060bb89",
    "metadata": {},
    "source": [
     "Print class and image of a single instance in the test data for preview."
@@ -143,7 +130,6 @@
   {
    "cell_type": "code",
    "execution_count": 4,
-   "id": "3828b316",
    "metadata": {},
    "outputs": [
     {
@@ -187,7 +173,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "ec223bf7",
    "metadata": {},
    "source": [
     "#### 2 - Compute and visualize the relevance attributions\n",
@@ -196,7 +181,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "c9eed64b",
    "metadata": {},
    "source": [
     "RISE masks random portions of the input image and passes the masked image through the model — the portion that decreases the accuracy the most is the most “important” portion.<br>\n",
@@ -206,7 +190,6 @@
   {
    "cell_type": "code",
    "execution_count": 5,
-   "id": "87d79611",
    "metadata": {},
    "outputs": [
     {
@@ -226,7 +209,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "305412e3",
    "metadata": {},
    "source": [
     "Visualize the relevance scores for the predicted class on top of the image."
@@ -235,7 +217,6 @@
   {
    "cell_type": "code",
    "execution_count": 7,
-   "id": "a61a1f43",
    "metadata": {},
    "outputs": [
     {
@@ -263,7 +244,6 @@
   },
   {
    "cell_type": "markdown",
-   "id": "4b661037",
    "metadata": {},
    "source": [
     "#### 3 - Conclusions\n",
@@ -275,7 +255,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
+   "display_name": "Python 3",
    "language": "python",
    "name": "python3"
   },
@@ -289,7 +269,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.7"
+   "version": "3.7.3"
   }
  },
  "nbformat": 4,

diff --git a/tutorials/explainers/RISE/rise_text.ipynb b/tutorials/explainers/RISE/rise_text.ipynb
@@ -9,9 +9,7 @@
     "### Interpreting a movie review sentiment model with RISE\n",
     "This notebook demonstrates the use of DIANNA with the RISE method on the [Stanford Sentiment Treebank dataset](https://nlp.stanford.edu/sentiment/index.html) which contains one-sentence movie reviews. See also [their paper](https://nlp.stanford.edu/~socherr/EMNLP2013_RNTN.pdf). A pre-trained neural network classifier is used, which identifies whether a movie review is positive or negative.\n",
     "\n",
-    "RISE is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates each word's relevance to the model's decision empirically by probing the model with randomly masked versions of the input image and obtaining the corresponding outputs.  \n",
-    "\n",
-    "More details about this method can be found in the paper https://arxiv.org/abs/1806.07421.\n",
+    "[RISE](http://bmvc2018.org/contents/papers/1064.pdf) is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates each word's relevance to the model's decision empirically by probing the model with randomly masked versions of the input image and obtaining the corresponding outputs.  \n",
     "\n",
     "*NOTE*: This tutorial is still work-in-progress, the final results need to be improved by tweaking the RISE parameters"
    ]

diff --git a/tutorials/explainers/RISE/rise_timeseries_frb.ipynb b/tutorials/explainers/RISE/rise_timeseries_frb.ipynb
@@ -13,9 +13,7 @@
     "### Model interpretation using RISE for astronomical timeseries data\n",
     "This notebook shows how to apply the RISE explainability method on a model trained to classify Fast Radio Burst (FRB) timeseries data. The model is a binary classifier that classifies the input as either noise or a real FRB.<br>\n",
     "\n",
-    "RISE is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates the relevance empirically by probing the model with randomly masked versions of the input image to obtain the corresponding outputs.<br>\n",
-    "\n",
-    "More details about this method can be found in the paper https://arxiv.org/abs/1806.07421.<br>"
+    "[RISE](http://bmvc2018.org/contents/papers/1064.pdf) is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates the relevance empirically by probing the model with randomly masked versions of the input image to obtain the corresponding outputs.<br>"
    ]
   },
   {

diff --git a/tutorials/explainers/RISE/rise_timeseries_weather.ipynb b/tutorials/explainers/RISE/rise_timeseries_weather.ipynb
@@ -17,9 +17,7 @@
     "\n",
     "It visualizes the relevance attributions for each segmentation of timeseries by displaying them on top of the timeseries.<br>\n",
     "\n",
-    "RISE is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates the relevance empirically by probing the model with randomly masked versions of the input image to obtain the corresponding outputs.<br>\n",
-    "\n",
-    "More details about this method can be found in the paper https://arxiv.org/abs/1806.07421.<br>"
+    "[RISE](http://bmvc2018.org/contents/papers/1064.pdf) is short for Randomized Input Sampling for Explanation of Black-box Models. It estimates the relevance empirically by probing the model with randomly masked versions of the input image to obtain the corresponding outputs.<br>"
    ]
   },
   {
@@ -1100,7 +1098,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
+   "display_name": "Python 3",
    "language": "python",
    "name": "python3"
   },
@@ -1114,7 +1112,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.7"
+   "version": "3.7.3"
   },
   "vscode": {
    "interpreter": {