Charles Du, Gagan Gupta, Michelle Nie, Alex Sands
AbstractWe have all had a hard time remembering specific details of movies we have watched or want to watch in the future. Oftentimes, we remember just the names of the main actors or actresses or genre of the movie. This creates a problem in that people want an easy way to communicate with a database to find information about movies, but traditional databases do not think like humans. In fact, we do not even speak the same language. To us, searching “show me movies about ______” or “show me movies featuring ______” is an intuitive way to find movies about something we are curious about or an actor we are interested in seeing. In SQL, these intuitive searches can become extremely long queries that are difficult for people who have not taken a databases course to write. Therefore, we wanted to simplify the search process by allowing people to speak to their computers.
We created a voice-enabled movie search web-app that displays results based on a user’s queries. Similar to Apple’s TV Remote interface or Comcast’s X1 Speech Remote interface, our application will help simplify currently complicated movie search. For example, if a user says, “Show me movies featuring Tom Cruise,” our application would return movies including Mission: Impossible III, Vanilla Sky, Knight and Day, and Valkyrie . talkie also provides recommendations for users. When users first log into the application, they will see the voice search interface and movie recommendations based on their rating history. Moreover, after a user rates a movie, we use a database of historical data to generate films the user might enjoy based on similar users’ preferences.