- Arduino/C++ implementation of Deep Policy Gradient Network with Reward-To-Go policy
- Custom built deep learning library for embedded controllers with floating point support
- Forward pass and backpropagation handled by means of custom matrix manipulation libraries written in C++
- Machine learning implementation close to popular libraries
- Built as final project for ENM 531: Data Driven Modelling and Probabilistic Scientific Computing course at UPenn
- Read more about it here
- Watch demo here.