-
-
-
-
A simple showcase of a conversation between user and BoomAI with the prompt of working in a hotel.
-
Experimental conversation with BoomAI using a generalized prompt that asks for the user's occupation and shows model's ability to translate.
Inspiration
Most language learning platforms give pre-planned lessons that always are uniform in style and are not custom to a user's needs.
What it does
BoomAI is a chatbot that aligns with the user's occupation and helps them learn phrases that are crucial to their daily lives. By specifically taking into consideration the user's occupation, BoomAI can simulate a conversation with the user using speech and offers input to their grammar skills and word choice.
How we built it
BoomAI utilizes the GPT model to generate responses based on a customized prompt and procedure that helps with BoomAI's general purpose. Using the Whisper API, speech is taken from the user and then taken in by the GPT model. We use Hume to invoke emotion in the sentences created by GPT that Azure carries out with the text-to-speech process.
Challenges we ran into
Understanding new APIs as well as building a front-end for the project.
Accomplishments that we're proud of
Addressing the main problem that we are trying to solve through the use of AI.
What we learned
New technologies such as Hume, Openai, and Azure.
What's next for BoomAI
- Improving generated responses both text and audio through an improved model that can convey emotion in other languages besides English.
- Integrating BoomAI's emotions in an observable way either through emojis or an actual avatar that depicts BoomAI.
Log in or sign up for Devpost to join the conversation.