Inspiration
Communicating freely with others in your society shouldn't be difficult, regardless of the language that each race use. Having a translator convert the spoken language to what everyone understands has been for many years, the pivotal solution to this barrier. Amongst societies, communities and peer groups, with the advent of technology, communication in different formalities should never be hindered as the medium of translation and interpretation need to be greatly improved.
What it does
Takes in audio in source language, gives text out in another.
How we built it
Greatly leveraged the Translator API with the Natural Language Pipeline built locally. Trained custom models and exposed them through a custom-built server for users to enjoy.
Challenges we ran into
- Time constraint (training models is expensive, use case must be clear).
- Interoperability (Models are embedded within chrome - that limits its reach to only google users).
- Processing power (our model's use case serves more than 10 tribes, and that equates to more processing power and use case).
Accomplishments that we're proud of
We just began building.
What we learned
We'd document it once we have concluded a maximum version.
What's next for GoogleAI
Extension of this powerful AI
Log in or sign up for Devpost to join the conversation.