Skip to content

alexandre-lavoie/deep-physics

Repository files navigation

Deep-Physics

Deep-Physics is a compact TensorFlow model trained to recreate intuitive physics. Deep-Physics allows for a simple generation of dynamic environments without the use of any classical physics formulas.

alt text

Usage

Deep-Physics can be applied to any environment. We chose to provide users with a simple javascript implementation using TensorFlow.js and P5.js. Simply clone the repository and run the index.html.

Research Paper

We have included our unpublished research paper which explains the theory behind our research.

Credits and Acknowledgment

Research project made by Alexandre Lavoie and James Kary from Vanier College, QC, Canada. We want to thank Dr. Ivan Ivanov for his guidance.

License

MIT License

About

Intuitive Physics of Dynamic Environments using DNNs, GANs, and RNNs

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages