Follow-up Question Recommendations for Knowledge Inquiry Systems
Term Project for CSE 583 - Pattern Recognition and Machine Learning Class
QA systems are build to answer questions, but sometimes the users don’t really know what they want to know; providing the right answer is not always feasible and/or enough.
The objective of this project: by leveraging machine learning techniques, a QA system can help build on the user’s knowledge and provide the user with relevant information even when no good answer is available.
In a sense, implement a recommender system for follow-up questions: suggest new questions to the user that might add to the already acquired knowledge.
Recommender systems are pretty common when there is a user profile available with stored information about user preferences, e.g. movie recommendations. In our case we have no preference data –we want to make recommendations by relying entirely on content, meaning the input question.