YosAI

Inspiration

Poems Schmoems, it's not that hard right? We're taking Artificial Intelligence together and thought this hackathon would be good practice for our group project.

What it does

YosAI utilizes Recurrent Neural Networks (RNN) and Natural Language Processing (NLP) to generate insightful haikus based off a robust dataset.

How we built it

We used Tensorflow and Keras for the frameworks to train our neural network. YosAI utilizes Long Short Term Memory (LSTM) via a Recurrent Neural Network (RNN) and Natural Language Processing (NLP) to generate insightful haikus based off a robust dataset.

The model uses character encoding and sequence prediction via batches to train in each epoch/generation.

Challenges we ran into

While being knowledgeable with AI theory, neither of us have had much practice with the frameworks necessary to create a meaningful predictive agent. This learning curve proved challenging as we applied what we learned. We also experimented with different methods of domain storing/encoding (words vs. syllables vs. characters) as well as predictions (words vs. characters). In the end, we decided to go with character encoding/prediction.

Accomplishments that we're proud of

We were able to create a functioning algorithm that can generate somewhat coherent haikus.

What we learned

The knowledge and experience gained from working on this project will be invaluable as we continue working on our AI group project over the coming weeks.