Final

Author: zacharyaanglin

0 - Introduction.ipynb

@@ -0,0 +1,142 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Bayesian Deep Learning in TensorFlow Probability and TensorFlow 2.0 -- The How and the Why"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    " "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    " "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "\n",
+    "    \n",
+    "![Hello](https://media1.giphy.com/media/ASd0Ukj0y3qMM/giphy.gif \"Wave\")\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## What this talk is\n",
+    "\n",
+    "- An overivew of the `layers` module of the recently-released TensorFlow Probability package and its integration into TensorFlow 2.0 `keras`\n",
+    "- An explanation of how to use `tfp.layers` in order to fit distributions over weights of neural networks and why you would want to do that\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    " \n",
+    " "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## What this talk isn't\n",
+    "- An introduction to probabilistic programming, Bayesian reasoning, and their subtopics in a general fashion\n",
+    "- A deep dive into the mechanics of variational inference and other techniques implemented in TensorFlow Probability\n",
+    "- An introduction to vanilla neural networks and their applications"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Who this talk is for\n",
+    "\n",
+    "- Pracitioners! Anyone interested in applying the latest in TensorFlow and TensorFlow Probability to solve interesting problems.\n",
+    "- People who are interested in marrying probabilistic techniques with deep learning frameworks, but haven't yet delved in themselves."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Why does this matter?"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/Users/anglin/.pyenv/versions/3.7.2/envs/tensorflow2/lib/python3.7/site-packages/IPython/core/display.py:701: UserWarning: Consider using IPython.display.IFrame instead\n",
+      "  warnings.warn(\"Consider using IPython.display.IFrame instead\")\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<iframe width=\"560\" height=\"315\" src=\"https://www.youtube.com/embed/HumFmLu3CJ8\" frameborder=\"0\" allow=\"accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture\" allowfullscreen></iframe>"
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from IPython.display import HTML\n",
+    "\n",
+    "HTML('<iframe width=\"560\" height=\"315\" src=\"https://www.youtube.com/embed/HumFmLu3CJ8\" frameborder=\"0\" allow=\"accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture\" allowfullscreen></iframe>')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Probabilistic deep learning is a field that's experienced hype in the recent past, but its depth of required knowledge and lack of high-level tools has discouraged many from being able to participate. Frameworks like Pyro on PyTorch have sought to address this problem, but so far uptake remains low.\n",
+    "\n",
+    "With the recent release of TensorFlow 2.0, the probabilistic learning framework TensorFlow Probability is included as a first class member of the TensorFlow ecosystem, and interoperates cleanly with TensorFlow 2.0 `keras`. This makes previous problems like fitting priors on the weights in a convolutional neural network layer much more approachable from a computational standpoint, allowing practitioners to focus on the theory and problems they are trying to solve."
+   ]
+  }
+ ],
+ "metadata": {
+  "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": "3.7.2"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}