homework/HW6/part1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# NLP Assignment 6\n",
    "Created by Prof. [Mohammad M. Ghassemi](https://ghassemi.xyz)\n",
    "\n",
    "Submitted by: <span style=\"color:red\"> INSERT YOUR NAME HERE </span>\n",
    "\n",
    "In collaboration with: <span style=\"color:red\"> INSERT YOUR (OPTIONAL) HOMEWORK PARTNER'S NAME HERE </span>\n",
    "\n",
    "<hr> "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Assignment Goals\n",
    "The goal of this assignment is to familiarize yourself with:\n",
    "\n",
    "1. Statistical Constituency Parsing\n",
    "\n",
    "The assignment combines tutorial components, with learning exercises that you must complete and submit. The learning exercise sections are clearly demarcated within the assignments.\n",
    "\n",
    "## Before you start\n",
    "1. PULL THE LATEST VERSION OF THE `course-materials` REPOSITORY, AND COPY `homework/HW6/` INTO THE CORRESPONDING DIRECTORY OF YOUR SUBMISSION FOLDER\n",
    "2. CREATE AND ATTACH TO A VIRTUAL ENVIRONMENT, AND INSTALL THE REQUIREMENTS IN `requirements.txt`\n",
    "3. IMPORT THE COURSE UTILITIES AND RELEVANT LIBRARIES BY RUNNING THE CODE BLOCK BELOW\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package brown to /Users/ghamut/nltk_data...\n",
      "[nltk_data]   Package brown is already up-to-date!\n",
      "[nltk_data] Downloading package universal_tagset to\n",
      "[nltk_data]     /Users/ghamut/nltk_data...\n",
      "[nltk_data]   Package universal_tagset is already up-to-date!\n",
      "[nltk_data] Downloading package treebank to /Users/ghamut/nltk_data...\n",
      "[nltk_data]   Package treebank is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import importlib\n",
    "from materials.code import utils\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import requests\n",
    "import nltk\n",
    "\n",
    "# IMPORT SOME BASIC TOOLS:\n",
    "from pprint import pprint\n",
    "from nltk.corpus import treebank\n",
    "import pyarrow\n",
    "\n",
    "from nltk.corpus import brown\n",
    "import ssl\n",
    "\n",
    "try:\n",
    "    _create_unverified_https_context = ssl._create_unverified_context\n",
    "except AttributeError:\n",
    "    pass\n",
    "else:\n",
    "    ssl._create_default_https_context = _create_unverified_https_context\n",
    "    \n",
    "nltk.download('brown')\n",
    "nltk.download('universal_tagset')\n",
    "nltk.download('treebank')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr> \n",
    "\n",
    "## Learning Exercise 1: \n",
    "\n",
    "<br>\n",
    "\n",
    "\n",
    "#### A. Data versus rule-driven approaches\n",
    "Describe a circumstance where you might expect a language model based on a Probabalistic Context Free Grammar to out-perform a purely data-driven approach.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<span style=\"color:red\"> INSERT AN INTERPRETATION OF YOUR RESULTS HERE </span>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "\n",
    "\n",
    "#### B. Language Generatation via Probabalistic Context Free Grammars\n",
    "Using the code below as a starting point, write a function that takes as an inputs (1) a lexicon, and (2) a probabalistic context free grammar and generates a sentence that adheres to the rules of the grammar. Your lexicon should be generated from the Brown corpus referenced in the code below. The probabilities for your grammar should come from the Penn Treebank, also referenced in the code below. Demonstrate your sentence generator with a Grammar that you specify, or extract from the Treebank."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<span style=\"color:red\"> INSERT AN INTERPRETATION OF YOUR RESULTS HERE </span>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br>\n",
    "\n",
    "\n",
    "#### C. Constituency Parsing with Probabilstic CKY\n",
    "Building on you work from part B, write a function that implements the probabilistic CKY algorithm to provide the most probable parse of a new sentence. Your function should take an input sentence, and a grammar, and return the most probable parse, and it's probability. Demonstrate your parser on the following sentences:\n",
    "\n",
    "1. \"My fat cat ate caviar\"\n",
    "2. \"My cat was fat and ate fish eggs\"\n",
    "3. \"Caviar was eaten by my fat cat\"\n",
    "4. \"The cat I own, which is fat, consumed fish eggs\"\n",
    "4. \"A cat that is fat, of which I am the owner, ate caviar\"\n",
    "\n",
    "Please recall that the the CKY algorithm assumes the grammar is in Chomsky Normal Form."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "################################################################################\n",
    "# INSERT YOUR CODE HERE\n",
    "# DO NOT FORGET TO PRINT YOUR MEANINGFUL RESULTS TO THE SCREEN.\n",
    "################################################################################"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<span style=\"color:red\"> INSERT AN INTERPRETATION OF YOUR RESULTS HERE </span>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "<h1><span style=\"color:red\"> Self Assessment </span></h1>\n",
    "Please provide an assessment of how successfully you accomplished the learning exercises in this assignment according to the instruction provided; do not assign yourself points for effort. This self assessment will be used as a starting point when I grade your assignments. Please note that if you over-estimate your grade on a given learning exercise, you will face a 50% penalty on the total points granted for that exercise. If you underestimate your grade, there will be no penalty.\n",
    "\n",
    "* Learning Exercise: \n",
    "    * A. <span style=\"color:red\">X</span>/1 points\n",
    "    * B. <span style=\"color:red\">X</span>/2 points\n",
    "    * C. <span style=\"color:red\">X</span>/2 points"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}