Since the late 1990s when speech companies began providing theircustomer-service software in the market, people have gotten used to speaking tomachines. As people interact more often with voice and gesture controlledmachines, they expect the machines to recognize different emotions, andunderstand other high level communication features such as humor, sarcasm andintention. In order to make such communication possible, the machines need anempathy module in them which can extract emotions from human speech andbehavior and can decide the correct response of the robot. Although research onempathetic robots is still in the early stage, we described our approach usingsignal processing techniques, sentiment analysis and machine learningalgorithms to make robots that can "understand" human emotion. We propose Zarathe Supergirl as a prototype system of empathetic robots. It is a softwarebased virtual android, with an animated cartoon character to present itself onthe screen. She will get "smarter" and more empathetic through its deeplearning algorithms, and by gathering more data and learning from it. In thispaper, we present our work so far in the areas of deep learning of emotion andsentiment recognition, as well as humor recognition. We hope to explore thefuture direction of android development and how it can help improve people'slives.
展开▼