{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Label Shuffle Experiment\n", "\n", "The progressive learning package utilizes representation ensembling algorithms to sequentially learn a representation for each task and ensemble both old and new representations for all future decisions. \n", "\n", "Here, a representation ensembling algorithm based on decision forests (Synergistic Forest) demonstrate forward and backward knowledge transfer of tasks on the CIFAR100 dataset with the labels shuffled. The experiment reproduces the benchmarking adversarial experiment ran in the paper \"Ensembling Representations for Synergistic Lifelong Learning with Quasilinear Complexity\" by Vogelstein, et al (2020). The following is a link to the aforementioned paper: https://arxiv.org/pdf/2004.12908.pdf \n", "\n", "### Import necessary packages and modules" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "from tensorflow import keras\n", "from joblib import Parallel, delayed\n", "from itertools import product" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Load CIFAR100 data \n", "We load the CIFAR100 dataset from Keras, and store it in a variable. The training and test partitions are concatenated into one variable called `data_x`. The data is obtained from https://keras.io/api/datasets/cifar100/ .\n", "\n", "The label shuffle experiment randomly permutes the class labels within each task from task 2 to 10, rendering each of these tasks adversarial with regard to the first task. We show through this experiment that SynF are invariant to class lable shuffling, and both demonstrate transfer. " ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "(X_train, y_train), (X_test, y_test) = keras.datasets.cifar100.load_data()\n", "data_x = np.concatenate([X_train, X_test])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Define hyperparameters for the model and preprocess data\n", "Running the cells below will define the hyperparameters the experimental setting \n", "\n", "`num_points_per_task`: The number of points per task \n", "\n", "`shifts`: The number of data ways to split the data into train and test" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "num_points_per_task = 500\n", "\n", "shifts = 2\n", "\n", "num_slots = int(5000 // num_points_per_task)\n", "slot_fold = range(int(5000 // num_points_per_task))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This cell will preprocess the data" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "# Reshape the data\n", "data_x = data_x.reshape(\n", " (data_x.shape[0], data_x.shape[1] * data_x.shape[2] * data_x.shape[3])\n", ")\n", "data_y = np.concatenate([y_train, y_test])\n", "data_y = data_y[:, 0]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Train the model and perform validation\n", "\n", "#### run_parallel_exp: \n", "Wrapper method for the `label_shuffle_experiment` function which declares and trains the model, and performs validation with respect to the test data to compute the error of the model at a particular iteration\n", "\n", "`ntree`: Number of trees for Uncertainty Forest" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "from functions.label_shuffle_functions import run_parallel_exp\n", "\n", "n_trees = [10] # Number of trees in SynF\n", "\n", "shift_fold = range(1, shifts, 1) # Number of shifts\n", "iterable = product(n_trees, shift_fold, slot_fold)\n", "\n", "df_list = Parallel(n_jobs=-1, verbose=0)(\n", " delayed(run_parallel_exp)(\n", " data_x, data_y, ntree, num_points_per_task, slot=slot, shift=shift\n", " )\n", " for ntree, shift, slot in iterable\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Function to calculate backward transfer efficiency\n", "\n", "The backward transfer efficiency of $f_n$ for task $t$ given $n$ samples is \n", "$$BTE^t (f_n) := \\mathbb{E} [R^t (f_n^{ 1$. Intuitively, this means that the progressive learner has used data associated with new tasks to improve performance on previous tasks. \n", "\n", "#### calc_bte:\n", "Function used to calculate bte across tasks, averaged across all shifts and folds" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "from functions.label_shuffle_functions import calc_bte\n", "\n", "btes = calc_bte(df_list, num_slots, shifts)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Plotting the backward transfer efficiency\n", "Run cell to generate plot of backward transfer efficiency of the Synergistic Forest algorithm. We see that we achieve backwards transfer overall that increases as more tasks are seen.\n", "\n", "#### plot_bte:\n", "Function used to plot bte across tasks" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "from functions.label_shuffle_functions import plot_bte\n", "\n", "plot_bte(btes)" ] } ], "metadata": { "interpreter": { "hash": "77d3befdf72f5c1a0d6b4996fdd6befdfb972b784410fca14e27e6ae1841315c" }, "kernelspec": { "display_name": "Python 3", "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": "3.6.9" } }, "nbformat": 4, "nbformat_minor": 2 }