Game World Building with Recurrent Neural Networks

Inspiration

The project was inspired by the rich worlds and characters found in popular literature and film. Special thanks to Gabriella Miesner for mentoring and guiding throughout the project.

What it does

The project utilizes recurrent neural networks to generate new character names based on existing ones.

How we built it

We built the project using TensorFlow and Keras. We created a vocabulary based on a sample of character names, converted the names into sequences of indices, and then used these sequences as input data for training the recurrent neural network.

Challenges we ran into

We encountered challenges in preprocessing the data and determining the appropriate architecture for the recurrent neural network.

Accomplishments that we're proud of

We're proud of successfully training the recurrent neural network to generate plausible character names.

What we learned

Through this project, we learned about sequence generation using recurrent neural networks and gained experience in working with text data.

What's next for Game World Building with Recurrent Neural Networks

In the future, we plan to explore more sophisticated architectures and train the model on larger datasets to further improve the quality of generated character names.

Built With

• python