AI-Powered Trivia Quiz Solution
Leveraging the Power of Knowledge, Mathematics, NLP and a Digital Platform
A POC was needed to test if algorithms could be successfully used to extract content about a particular topic from existing data sources, prepare a questionnaire based on it and present it to users as a quiz game.
We built an interesting trivia game using slack integration and cognitive services. The concept was to use news API to find relevant news from the topic chosen by the user.
- For the chosen topic, based on the latest news headlines, cognitive services were used to find relevant news headlines that contain People, Places, or Companies. The Matching headlines were then converted into multiple-choice questions. NLP and custom algorithms were used to build alternate choices for the Multiple choice questions.
- BOT frameworks were used to present the quiz to users via various channels like Slack, MS teams, Skype and Facebook chat. Users will provide topics and Bot will decide questions to ask, find interesting alternate answers and randomize relevant answers as multiple options to users. Question selection is completely random yet relevant.
- Leaderboard, timer, etc. are some other great features built within the BOT.
An alternate version of the trivia quiz used the Wikipedia page as the data source for a user-given keyword. A tokenizer algorithm was used to extract sentences from the Wikipedia article, recognize the most significant word in that sentence, and replace it with a blank space. The remaining three options were generated using Wikidata and converted into a fill-in-the-blanks multiple choice quiz question.
ASP.NET MVC, Slack, Slack integration, Slackbot , Microsoft BOT framework,
Azure Cognitive services, Microsoft NLP, Google NLP,
Wikipedia API, Slack integration, Microsoft Teams integration, Node js, Tokenizer for Python, Azure Table storage, News API
- Successfully executed the POC project to extract content from a knowledge base and create a multiple choice quiz based on it
- This concept proved to have promising commercial implications in the education domain (to generate questions from a book) and as a training tool in companies