This project aims at investigating the construction of humor models to enrich conversational agents through the help of interactive reinforcement learning approaches.
Our methodology consisted in deploying an online platform where passersby can play a game of matching cards in order to generate humorous sentences. This game was inspired by popular similar card games.
The data collected from these interactions allowed us to start investigating how humour emerges in semantic associations. 
We worked on this project for 4 months, resulting in an implementation of a prototype which includes a first humor-enabled embodied conversational agent.