Better example plots #13

public/images/bandit-learning-gp-GPy-Copy1.ipynb → public/images/main.ipynb

@@ -486,7 +486,7 @@
        "</table>"
       ],
       "text/plain": [
-       "<GPy.models.gp_regression.GPRegression at 0x7f14088b4f60>"
+       "<GPy.models.gp_regression.GPRegression at 0x7f96eb4de668>"
       ]
      },
      "execution_count": 15,
@@ -1781,36 +1781,6 @@
     "        return pi_acq + lam * ei_acq"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": 56,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# acq_params = {\n",
-    "#     'eps_pi': .2,\n",
-    "#     'eps_ei': .4,\n",
-    "#     'lam': 1,\n",
-    "# }\n",
-    "\n",
-    "# dirName, mtn_ei_pi = plot_acquisition(\n",
-    "#     train_X, train_y,\n",
-    "#     EI_PI, val, ylim1=((0, 10)),\n",
-    "#     acq_params=acq_params,\n",
-    "#     y1scale='log',\n",
-    "# )\n",
-    "\n",
-    "# # gify!\n",
-    "# !convert -delay {delay} -loop 0 {dirName}/*.png {gifDir}/mab-gp-ei_pi.gif"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "![](MAB_gifs/mab-gp-ei_pi.gif)"
-   ]
-  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -1824,7 +1794,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 57,
+   "execution_count": 56,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1833,7 +1803,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 58,
+   "execution_count": 57,
    "metadata": {
     "scrolled": false
    },
@@ -1967,7 +1937,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 59,
+   "execution_count": 58,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -1979,7 +1949,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 60,
+   "execution_count": 59,
    "metadata": {},
    "outputs": [
     {
@@ -2016,7 +1986,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 61,
+   "execution_count": 60,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2042,7 +2012,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 62,
+   "execution_count": 61,
    "metadata": {
     "scrolled": true
    },
@@ -2073,7 +2043,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 63,
+   "execution_count": 62,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2083,7 +2053,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 64,
+   "execution_count": 63,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2097,18 +2067,36 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 65,
+   "execution_count": 64,
    "metadata": {},
    "outputs": [],
    "source": [
     "nnsvm(plt.rcParams, 4)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 65,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from matplotlib.markers import MarkerStyle"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 66,
    "metadata": {},
    "outputs": [],
+   "source": [
+    "plt.scatter?"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 67,
+   "metadata": {},
+   "outputs": [],
    "source": [
     "# https://matplotlib.org/3.1.1/gallery/ticks_and_spines/colorbar_tick_labelling_demo.html\n",
     "import sys\n",
@@ -2209,16 +2197,18 @@
     "        ax[0].scatter(\n",
     "            train_X[:, 0],\n",
     "            train_X[:, 1],\n",
-    "            color='plum',\n",
-    "            s=30,\n",
+    "            facecolors='none', \n",
+    "            edgecolors='k',\n",
+    "            linewidth=4,\n",
+    "            s=100,\n",
     "            zorder=10,\n",
     "            label='Training points'\n",
     "        )\n",
     "        # plot the global maxima\n",
     "        ax[0].scatter(\n",
     "            *max_x,\n",
     "            marker=\"*\", \n",
-    "            color='yellow',\n",
+    "            color='gold',\n",
     "            s=300,\n",
     "            zorder=20,\n",
     "            label=\"Gobal Maxima\"\n",
@@ -2239,40 +2229,19 @@
     "            ticks=ticks2,\n",
     "            ax=ax[1]\n",
     "        )\n",
-    "        \n",
-    "        if str(acq_obj) == \"PI\":\n",
-    "            ax[1].set_title(\n",
-    "                r'$\\alpha_{PI}$' + f\"\\nIteration: {i} \\n\" + \\\n",
-    "                r\"$\\epsilon$ = \" + str(acq_params['eps'])\n",
-    "            )\n",
-    "        elif str(acq_obj) == \"EI\":\n",
-    "            ax[1].set_title(\n",
-    "                r'$\\alpha_{EI}$' + f\"\\nIteration: {i} \\n\" + \\\n",
-    "                r\"$\\epsilon$ = \" + str(acq_params['eps'])\n",
-    "            )\n",
-    "        elif str(acq_obj) == \"GP_UCB\":\n",
-    "            ax[1].set_title(\n",
-    "                r'$\\alpha_{GP-UCB}$' + f\"\\nIteration: {i} \\n\" + \\\n",
-    "                r\"$v$ = \" + str(acq_params['v']) + '\\n' \\\n",
-    "                r\"$\\delta$ = \" + str(acq_params['delta'])\n",
-    "            )\n",
-    "        elif str(acq_obj) == \"Rand\":\n",
-    "            ax[1].set_title(\n",
-    "                r'$\\alpha_{Random}$' + f\"\\nIteration: {i}\"\n",
-    "            )\n",
-    "        else:\n",
-    "            raise ValueError\n",
+    "        ax[1].set_title(f\"Acquisition Function ({str(acq_obj)})\")\n",
+    "        fig.suptitle(f\"Iteration: {i}\", fontsize='x-large')\n",
     "            \n",
     "        if svm:\n",
-    "            ax[1].set_xlabel(r\"$\\log_{10}\\gamma$\")\n",
-    "            ax[1].set_ylabel(r\"$C$\")\n",
-    "            ax[0].set_xlabel(r\"$\\log_{10}\\gamma$\")\n",
-    "            ax[0].set_ylabel(r\"$C$\")\n",
+    "            ax[1].set_xlabel(r\"Hyperparam #1 $(\\gamma)$\")\n",
+    "            ax[1].set_ylabel(r\"Hyperparam #2 $(C)$\")\n",
+    "            ax[0].set_xlabel(r\"Hyperparam #1 $(\\gamma)$\")\n",
+    "            ax[0].set_ylabel(r\"Hyperparam #2 $(C)$\")\n",
     "        else:\n",
-    "            ax[1].set_xlabel(\"# of Trees\")\n",
-    "            ax[1].set_ylabel(\"Max Depth\")\n",
-    "            ax[0].set_xlabel(\"# of Trees\")\n",
-    "            ax[0].set_ylabel(\"Max Depth\")\n",
+    "            ax[1].set_xlabel(\"Hyperparam #1 (# of Trees)\")\n",
+    "            ax[1].set_ylabel(\"Hyperparam #2 (Max Depth)\")\n",
+    "            ax[0].set_xlabel(\"Hyperparam #1 (# of Trees)\")\n",
+    "            ax[0].set_ylabel(\"Hyperparam #2 (Max Depth)\")\n",
     "\n",
     "        # Add new point with highest uncertainty to the pool set\n",
     "        ax[1].scatter(\n",
@@ -2327,7 +2296,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 67,
+   "execution_count": 68,
    "metadata": {
     "scrolled": false
    },
@@ -2357,7 +2326,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 68,
+   "execution_count": 69,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2384,7 +2353,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 69,
+   "execution_count": 70,
    "metadata": {
     "scrolled": true
    },
@@ -2418,7 +2387,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 70,
+   "execution_count": 71,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2451,7 +2420,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 71,
+   "execution_count": 72,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2500,7 +2469,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 72,
+   "execution_count": 73,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2691,7 +2660,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 73,
+   "execution_count": 74,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2700,7 +2669,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 74,
+   "execution_count": 75,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2726,7 +2695,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 75,
+   "execution_count": 76,
    "metadata": {
     "scrolled": true
    },
@@ -2757,7 +2726,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 76,
+   "execution_count": 77,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2766,7 +2735,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 77,
+   "execution_count": 78,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2906,7 +2875,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 78,
+   "execution_count": 79,
    "metadata": {
     "scrolled": false
    },
@@ -2936,7 +2905,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 79,
+   "execution_count": 80,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -2964,7 +2933,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 80,
+   "execution_count": 81,
    "metadata": {
     "scrolled": true
    },
@@ -2998,7 +2967,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 81,
+   "execution_count": 82,
    "metadata": {
     "scrolled": false
    },
@@ -3034,7 +3003,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 82,
+   "execution_count": 83,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -3043,7 +3012,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 83,
+   "execution_count": 84,
    "metadata": {},
    "outputs": [
     {