Chatmobile: Personality-driven chatbot over mobile

Abstract

The notion of being able to communicate with artificially intelligent entities in natural language is certainly not a new one, and can be seen in various works of science fiction. Today, natural language understanding is still an active area of research. This project explores chatbots, which are a particular type of dialogue system usually aimed at entertaining the user by engaging in conversation and attempting to appear somewhat human by generating responses that show some understanding of the input utterances. ELIZA, the first chatbot, was created by Joseph Weizenbaum in 1966. Chatbots rose to popularity in the 90s, partly because of the availability of technologies that simplified the development of ELIZA-like chatbots. However, a principal contributing factor to their popularity was the advent of the Loebner Prize competition, a contest based on a setting that resembles A. M. Turing's imitation game, which has dominated the chatbot scene in recent years. The Loebner Prize competition has been criticised for attracting more media attention than promotion of scientific research. In fact, to this day, Loebner Prize contestants and winners still use techniques similar to those used by ELIZA back in 1966. This project aims to break away from the trend by using natural language processing and linguistic techniques for natural language understanding, rather than keyword and pattern matching. The artefact created for this project also incorporates elements of learning, question answering, context handling, personality and mobile communications, all of which are usually either inexistent or very limited in other chatbots.