QA Generation
ICT & Artificial Intelligence
Client company:Infoland
Joey Lau
Brian Onuoha
Hendry Lloyd
Dylan Arts
Project description
How can Dutch multiple-choice questions be generated using AI?
Context
Generating questions and answers is a tedious task. Our client, Infoland, has offered us a project focused on generating multiple questions answers from a given text using artificial intelligence. This project will be part of their service Zenya Boost, to help optimize the employee training process.
Results
We created an (AI) Artificial Intelligence with deep learning that is able to interpret a given context and produce a quality question based on a topic that resides in the context. This question is then also answered by a separate AI that determines the correct answer to the question based on both the question and context. After generating the correct answer we have a proof of concept AI model that creates distractor answers. These answers are meant to look like they could be correct but aren't. We all wrapped this in a simple demo project to show how it all works together.