Inspiration
Articulate was inspired by the lack of available language learning apps that truly center around their users by uniquely leveraging AI to target pain points in language pronunciation. We all had something in common, we all came from a family that spoke more than one language, and from there we ideated Articulate.
What it does
Articulate works like a basic language learning app, however it emphasizes conversation by leveraging AI with iterative learning, which allows users to gain instant feedback on their pronunciations. These feedback reports are personalized on a level that maximizes user efforts to gain practical phonetic skills in the languages they desire.
How we built it
After ideation, the UI UX team commenced research on competitors to locate pain points in common language learning apps. From there, the entire team brainstormed additional features and drew out the functional and non-functional requirements. The UI UX team created the priority frames for the application to pass off to the development team, who during the time was also researching OpenAI documentations, setting up the environment using Python, code integrations in terms of api client, authentications, functionality implementations, and Training the model using Mozilla Common Voice.
Challenges we ran into
We ran into challenges with the rate limitations on OpenAI and Training the model to certain levels of accuracy.
What we learned
We learned a great deal on OpenAI API functionality implementations in terms of handling responses, calls, and processing results. We also learned the importance of research and truly understanding holistic user perspectives.
Accomplishments that we're proud of
We are proud that we got the opportunity to work together as a team to learn from each other and to create something that is meaningful. The principle of unity is something that the world should strive for, and a path to that is through communication, language, resilience, education, and understanding, just a few of the important aspects that Articulate represents.
What's next for Articulate
Articulate would like to have a more refined result system. Currently we offer a percentage of accuracy, followed by personalized exercises and feedback, however additional criteria and stats would help user connection with the app. Implementations for grammar is something that is feasible that we are interested in after our main focus in phonetics. Now, in terms of phonetic support methods, we have a small handful already that are effective, but there are a multitude of methods out there. Commencing additional research to implement these strategies will be proven as worthwhile endeavors in the ever changing landscape of education, language, and communication.
Built With
- css
- figma
- flask
- html
- openai
- python
- react-native
Log in or sign up for Devpost to join the conversation.