Adaptive Dialog Techniques for Chatbots

This research focuses on the development of machine learning and natural language processing
techniques for adaptive chatbots that can converse with users and learn about the world. The
outcome will be a generic bot that can be instantiated with different personalities and knowledge
bases to entertain users of the popular Kik messenger about different topics. It will improve the
dialog abilities of Kik’s current chat bot, which receives more than 1 million messages per day.
The immediate benefit will be higher user engagement. The second benefit will be a proof of
concept for future chatbots that celebrities may use to entertain their fans or companies may use
for one-on-one interactions with their customers. This work will propel Canada at the forefront
of automated dialog systems, which are seen as the key for improved user experience in
smartphones, personalized marketing and video-gaming.
Mitacs

Faculty Supervisor:

Pascal Poupart;Jesse Hoey

Student:

Partner:

KIK Interactive

Discipline:

Computer science

Sector:

Information and cultural industries

University:

University of Waterloo

Program: