{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Unsupervised Learning\n", "\n", "In unsupervised learning, labels are either not available or are not used for learning. Clustering is a form of unsupervised learning whereby, without labels, the data is characterized by constraints alone.\n", "LTN can formulate such constraints, such as for example: \n", "- clusters should be disjoint,\n", "- every example should be assigned to a cluster,\n", "- a cluster should not be empty,\n", "- if the points are near, they should belong to the same cluster,\n", "- if the points are far, they should belong to different clusters." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Init Plugin\n", "Init Graph Optimizer\n", "Init Kernel\n" ] } ], "source": [ "import logging; logging.basicConfig(level=logging.INFO)\n", "import tensorflow as tf\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "import ltn" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Data" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# loading data\n", "\n", "nr_of_clusters = 4\n", "nr_of_points_x_cluster = 50\n", "clst_ids = range(nr_of_clusters)\n", "\n", "close_threshold = 0.2\n", "distant_threshold = 1.0\n", "\n", "margin = .2\n", "mean = [np.random.uniform([-1+margin,-1+margin],[0-margin,0-margin],2),\n", " np.random.uniform([0+margin,-1+margin],[1-margin,0-margin],2),\n", " np.random.uniform([-1+margin,0+margin],[0-margin,1-margin],2),\n", " np.random.uniform([0+margin,0+margin],[1-margin,1-margin],2)]\n", "\n", "cov = np.array([[[.01,0],[0,.01]]]*nr_of_clusters)\n", "\n", "cluster_data = {}\n", "for i in clst_ids:\n", " cluster_data[i] = np.random.multivariate_normal(mean=mean[i],cov=cov[i],size=nr_of_points_x_cluster)\n", "\n", "data = np.concatenate([cluster_data[i] for i in clst_ids])\n", "\n", "for i in clst_ids:\n", " plt.scatter(cluster_data[i][:, 0], cluster_data[i][:, 1])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Language" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "Not = ltn.Wrapper_Connective(ltn.fuzzy_ops.Not_Std())\n", "And = ltn.Wrapper_Connective(ltn.fuzzy_ops.And_Prod())\n", "Or = ltn.Wrapper_Connective(ltn.fuzzy_ops.Or_ProbSum())\n", "Implies = ltn.Wrapper_Connective(ltn.fuzzy_ops.Implies_Reichenbach())\n", "Equiv = ltn.Wrapper_Connective(ltn.fuzzy_ops.Equiv(ltn.fuzzy_ops.And_Prod(),ltn.fuzzy_ops.Implies_Reichenbach()))\n", "Forall = ltn.Wrapper_Quantifier(ltn.fuzzy_ops.Aggreg_pMeanError(p=4),semantics=\"forall\")\n", "Exists = ltn.Wrapper_Quantifier(ltn.fuzzy_ops.Aggreg_pMean(p=6),semantics=\"exists\")" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Metal device set to: Apple M1\n", "\n", "systemMemory: 16.00 GB\n", "maxCacheSize: 5.33 GB\n", "\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "2021-10-15 18:53:14.098343: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:305] Could not identify NUMA node of platform GPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.\n", "2021-10-15 18:53:14.098434: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:271] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: METAL, pci bus id: )\n" ] } ], "source": [ "from tensorflow.keras import layers\n", "\n", "class MLP_classifier(tf.keras.Model):\n", " \"\"\" Model to call as P(x,class) \"\"\"\n", " def __init__(self, n_classes, single_label, hidden_layer_sizes=(16,16,16)):\n", " super(MLP_classifier, self).__init__()\n", " self.denses = [layers.Dense(s, activation=\"elu\") for s in hidden_layer_sizes]\n", " self.dense_class = layers.Dense(n_classes)\n", " self.to_probs = tf.nn.softmax if single_label else tf.math.sigmoid\n", " \n", " def call(self, inputs):\n", " x, c = inputs[0], inputs[1]\n", " for dense in self.denses:\n", " x = dense(x)\n", " logits = self.dense_class(x)\n", " probs = self.to_probs(logits)\n", " indices = tf.cast(c, tf.int32)\n", " return tf.gather(probs, indices, batch_dims=1) \n", "\n", "C = ltn.Predicate(MLP_classifier(nr_of_clusters, single_label=True))\n", "cluster = ltn.Variable(\"cluster\",clst_ids)\n", "\n", "x = ltn.Variable(\"x\",data)\n", "y = ltn.Variable(\"y\",data)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Notice that we use a clustering predicate that ends in a `softmax` layer (`single_label=True`), which returns mutually-exclusive probabilities for each cluster.\n", "Therefore, there is no explicit constraint in the knowledgebase about clusters being disjoint. \n", "\n", "Notice also the use of guarded quantifiers in the two last axioms: all the pairs of points with Euclidean distance lower (resp. higher) than some threshold should belong in the same cluster (resp.should not). The thresholds (`0.2` and `1.0` in this case) do not have to explicitly define all existing pairs of points as \"close\" or \"distant\"; they only define *some of* the closest and most distant pairs of points." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "formula_aggregator = ltn.Wrapper_Formula_Aggregator(ltn.fuzzy_ops.Aggreg_pMeanError(p=2))\n", "\n", "eucl_dist = ltn.Function.Lambda(lambda inputs: tf.expand_dims(tf.norm(inputs[0]-inputs[1],axis=1),axis=1))\n", "is_greater_than = ltn.Predicate.Lambda(lambda inputs: inputs[0] > inputs[1])\n", "\n", "close_thr = ltn.Constant(close_threshold, trainable=False)\n", "distant_thr = ltn.Constant(distant_threshold, trainable=False)\n", "\n", "# defining the theory\n", "#@tf.function\n", "def axioms(p_exists):\n", " axioms = [\n", " Forall(x, Exists(cluster, C([x,cluster]),p=p_exists)),\n", " Forall(cluster, Exists(x, C([x,cluster]),p=p_exists)),\n", " Forall([cluster,x,y], Equiv(C([x,cluster]),C([y,cluster])),\n", " mask = is_greater_than([close_thr,eucl_dist([x,y])])),\n", " Forall([cluster,x,y], Not(And(C([x,cluster]),C([y,cluster]))),\n", " mask = is_greater_than([eucl_dist([x,y]),distant_thr]))\n", " ]\n", " sat_level = formula_aggregator(axioms).tensor\n", " return sat_level\n" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# first call to build the graph\n", "axioms(p_exists=6)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Training" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 0: Sat Level 0.441\n", "Epoch 100: Sat Level 0.442\n", "Epoch 200: Sat Level 0.641\n", "Epoch 300: Sat Level 0.849\n", "Epoch 400: Sat Level 0.852\n", "Epoch 500: Sat Level 0.853\n", "Epoch 600: Sat Level 0.853\n", "Epoch 700: Sat Level 0.853\n", "Epoch 800: Sat Level 0.854\n", "Epoch 900: Sat Level 0.854\n", "Training finished at Epoch 999 with Sat Level 0.854\n" ] } ], "source": [ "trainable_variables = C.trainable_variables\n", "optimizer = tf.keras.optimizers.Adam(learning_rate=0.001)\n", "\n", "for epoch in range(1000):\n", " if epoch <= 100:\n", " p_exists = 1\n", " else:\n", " p_exists = 6\n", " with tf.GradientTape() as tape:\n", " loss_value = 1. - axioms(p_exists=p_exists)\n", " grads = tape.gradient(loss_value, trainable_variables)\n", " optimizer.apply_gradients(zip(grads, trainable_variables))\n", " if epoch%100 == 0:\n", " print(\"Epoch %d: Sat Level %.3f\"%(epoch, axioms(p_exists=p_exists)))\n", "print(\"Training finished at Epoch %d with Sat Level %.3f\"%(epoch, axioms(p_exists=p_exists)))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Print result" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "plt.rcParams['font.size'] = 12" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "x0 = data[:, 0]\n", "x1 = data[:, 1]\n", "\n", "prC = [C.model([data, tf.constant([[i]]*len(data))]) for i in clst_ids]\n", "n = 2\n", "m = (nr_of_clusters + 1) // n + 1\n", "\n", "fig = plt.figure(figsize=(10, m * 3))\n", "\n", "plt.subplots_adjust(wspace=0.2,hspace=0.3)\n", "ax = plt.subplot2grid((m,8),(0,2),colspan=4)\n", "ax.set_title(\"groundtruth\")\n", "for i in clst_ids:\n", " ax.scatter(cluster_data[i][:, 0], cluster_data[i][:, 1])\n", "for i in clst_ids:\n", " fig.add_subplot(m, n, i + 3)\n", " plt.title(\"C\" + str(i) + \"(x)\")\n", " plt.scatter(x0, x1, c=prC[i],vmin=0,vmax=1)\n", " plt.colorbar()\n", "# plt.savefig(\"ex_clustering_test.pdf\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "interpreter": { "hash": "12eaedf9b9a64329743e8900a3192e3d75dbaaa78715534825922e4a4f7d9137" }, "kernelspec": { "display_name": "Python 3.9.6 64-bit ('tf-py39': conda)", "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": "3.9.6" } }, "nbformat": 4, "nbformat_minor": 4 }