-
Sample image of a conversation with AWS version of PatriotBot. (3)
-
All photos are examples of how the Amazon Lex Patriot Bot looked like. This is a screenshot of the control flow.
-
Sample image of a conversation with AWS version of PatriotBot. (2)
-
Sample image of a conversation with AWS version of PatriotBot. (1)
Inspiration
The reason we created PatriotBot was because as incoming freshman, navigating the GMU website and finding information that was relevant to us was difficult, and we wanted to solve this problem for future incoming students.
What it does
PatriotBot is a chatbot website that helps incoming freshman navigate their college experience much more seamlessly, saving them the hassle of sifting through the entire GMU website themselves.
How we built it
We originally built a MVP using Amazon Lex (AWS), but then transferred over to Python, and used that, Flask, NLTK, and APIs like OpenAI to achieve a more functional product.
Challenges we ran into
Initially, we were planning on creating our chatbot through Amazon Lex. We were able to create a functioning chatbot, but we felt, as beginners on AWS, the conditional branching needed to allow our desired amount of possible user inputs was complicated to implement, so we used Python instead, to give us full control over our program.
Accomplishments that we're proud of
We're proud that we created PatriotBot, and learned other technologies along the way such as Amazon Lex and AWS Lambda, and more about Python and incorporating APIs. We also feel that a service like this would be very useful to Mason students, or college students in general, and we are excited to learn about the many possible opportunities to implement a program like this in universities.
What we learned
We learned our way around AWS and its use cases, and we learned about the many techniques used in making chatbots, web scraping, conversational AI, how ML models judge the similarity of texts, and more.
What's next for PatriotBot
We want to expand this project to allow for more information to be stored, allowing for a wider range of topics the bot can answer questions on. We'd also want to explore turning this into a service that other colleges, not just GMU, can use to help their newer students as well.
Log in or sign up for Devpost to join the conversation.