In 1966, Joseph Weizenbaum published an article about his chatbot ELIZA, called ELIZA—a computer program for the study of natural language communication between man and machine. Created with a sense of humor to show the limitations of technology, the chatbot employed simplistic rules and vague, open-ended questions as a way of giving an impression of empathic understanding in the conversation, and was in an ironic twist often seen as a milestone of artificial intelligence. The field has moved on, and today, AI assistants are around us: you might have an Alexa, a Google Echo, or any of the other commercial home assistants in the market.
In this recipe, we'll be building an AI assistant. The difficulty with this is that there are an infinite amount of ways for people to express themselves and that it is simply impossible to anticipate everything your users might say. In this recipe, we'll be training a model to infer what they want and we'll respond...