Description
Quiz games let people test their knowledge through multi-choice questions. Unfortunately, generating such questions can be very time-consuming and it is typically done manually. In this talk we will present a pipeline to automatically generate quiz games starting from generic knowledge (e.g. Wikipedia). The pipeline consists of the following components: (i) a parser to retrieve text from Wikipedia pages, (ii) a Natural Language Processing module (based on the Google Natural Language API) to extract information about syntax, entities and relations, (iii) a Natural Language Generation module to generate test questions and correct answers, and finally (iv) a domain-aware module that uses domain-specific knowledge to generate wrong answers (i.e. distractors). Every module is written in Python and it is based on either available libraries or Cloud services (e.g., Google Natural Language).
Prerequisites: basic knowledge of python, minimal understanding of text syntactic analysis.