Coreference Perturbations Test.ipynb

{
 "cells": [
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "// Turby: data augmentation for text\nimport com.digitalreasoning.tda.*\n\n// Sample coreference texts\n// From: https://en.wikipedia.org/wiki/Coreference#Types_of_coreference\nval texts: List<String> = listOf(\n    \"Every farmer who owns a donkey beats it.\",\n    \"Every police officer who arrested a murderer insulted him.\"\n)\n\n// Initialize and configure coreference perturber\nval coreferencePerturber = CoreferencePerturber\ncoreferencePerturber.verbose = true\n\n// Get perturbations\ntexts.forEachIndexed { index, text ->\n    println(\"Example ${index + 1}:\")\n    val perturbations = coreferencePerturber.perturb(text)\n    println()\n}",
   "execution_count": 1,
   "outputs": [
    {
     "output_type": "stream",
     "text": "Example 1:\n\nOriginal sentences:\n(0): `Every farmer who owns a donkey beats it.`\n\nCoreference chains:\nCHAIN0-[\"Every farmer who owns a donkey\" in sentence 1]\nCHAIN2-[\"a donkey\" in sentence 1, \"it\" in sentence 1]\n\nPerturbed sentences:\nStrategy: transform mentions to representative mention\n(0): `Every farmer who owns a donkey beats a donkey.`\nStrategy: transform mentions to pronominal form\n(0): `He beats it.`\n(0): `Every farmer who owns it beats it.`\n\nExample 2:\n\nOriginal sentences:\n(0): `Every police officer who arrested a murderer insulted him.`\n\nCoreference chains:\nCHAIN0-[\"Every police officer who arrested a murderer\" in sentence 1]\nCHAIN2-[\"a murderer\" in sentence 1, \"him\" in sentence 1]\n\nPerturbed sentences:\nStrategy: transform mentions to representative mention\n(0): `Every police officer who arrested a murderer insulted a murderer.`\nStrategy: transform mentions to pronominal form\n(0): `He insulted him.`\n(0): `Every police officer who arrested him insulted him.`\n\n",
     "name": "stdout"
    }
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "kotlin",
   "display_name": "Kotlin",
   "language": "kotlin"
  },
  "language_info": {
   "file_extension": "kt",
   "name": "kotlin"
  },
  "gist_id": "c540928da85f6dbc43d7b5324ec64010"
 },
 "nbformat": 4,
 "nbformat_minor": 1
}

Coreference Perturbations.ipynb