Say a student wanted to search for the formula for a Fourier transform on Piazza, but gets 40 posts relating to Fourier transform. Another student is looking for a particular topic and wants someone to guide them step-by-step to the response they are looking for. A third student then spends plenty of time looking for answers on the web, but finds out that the discussion board his organization or course uses has already discussed it. Lastly, a professor wants an easy way to send emails to students to update them on a specific topic. Oh, if only there were an easier way to deal with all this!
This is the aim of ChattyBuddy: to build an online ChatBot for discussion platforms to aid its users in finding the answers they seek. Perhaps the answer they are looking for is already on the board, but merely difficult to find. Or perhaps the student needs a study partner and can get him or herself a 24-hour virtual study buddy in the form of a chatbot. Or they may just wanna goof off with someone. In any case, our chatbot is designed for the purpose of maximizing client, user, student, and consumer satisfaction when it comes to more convenient answer seeking.
Ashwin is a senior Electrical Engineer with a Computer Design focus at UCSD. Amidst solving interesting computer science challenges, he enjoys playing basketball, snorkeling with leopard sharks and learning to surf.
Jerry is a senior Electrical Engineer with a Computer Systems and Data Communications focus at UCSD. Aside from designing and writing programs, and building circuits, he likes to play chess and run on treadmills.
Po is a junior Electrical Engineer with a focus in Machine Learning at UCSD. He has worked hard to build up his technical skills and wants to apply them to help better people's lives. In his free time, he likes to play board games, watch shows, and catch up on the sleep he never gets.
Tommy is a junior Electrical Engineer with a focus in Machine Learning at UCSD.In addition to interest in programming and robotics, he also enjoys outdoor activities and watching sitcoms.
After conducting some research on how to build a chatbot, we decided to implement it by using python and its deep learning module called PyTorch. Using a dataset of common Discussion Board questions and responses, as well as some conversational data, we trained our Chatbot to determine appropriate responses to any question or statement given to it, including the statement type and the appropriate response, be it a search on the discussion board for an answer, or a witty remark.
Our implementation still has a few bugs to crush. We will continue updating it to ensure its functionality.