Inspiration: Salvage Cal students from the burden of managing their heavy load of course works.
What Calbot Does: Calbot is an artificial-intelligent chatting bot on Facebook messenger that assists UC Berkeley students to plan their college lives. Main functions and features include enrolling and dropping classes, adding class assignments, querying a list of user’s classes info and due dates, natural language processing, etc.
How we built it: The main framework was written in JavaScript, and our backend server was built on Amazon Web Service. We’ve also used IBM Watson cognitive system for training the bot.
Challenges we ran into: Reconciling Watson with the bot framework; training Watson to make the bot intelligence and reply relevance ideal.
Accomplishments that we’re proud of: Training our bot to an ideal level of intelligence; building a stable framework for accessing and modifying the database.
What we learned: Using the Facebook messenger API, training a bot with IBM Watson’s machine learning algorithm, using database appropriately.
What’s next for Calbot: In-picture object recognition feature; comprehensively and maturely trained AI.
Log in or sign up for Devpost to join the conversation.