Visualizing the Steering Model with Attention Maps

Introduction Convolutional neural networks are often known as “black boxes” for their mysterious nature. Unlike most programs that we write, computer scientist can’t directly modify the content (weights) of the neural networks to improve their performance. In order to create better machine learning models, you can either do a heck more training or experiments with […]

Read more

The Limitations Of Our Deep Learning Powered Self-Driving Golf Cart

Introduction After our somewhat unsuccessful demo last Wednesday, my partner Michael Meng exclaimed, “there is no hope for deep learning”. The future is not that grim, but Michael is right to a certain degree. Deep learning has flaws and our deep learning powered self-driving golf cart certainly has lots of flaws. Today, I would like to […]

Read more

The Robustness of the Semantic Segmentation Network

Introduction On Feb 21st, Michael and I tested the autonomous steering and cruise control system. Unfortunately, the testing was largely unsuccessful. We encounter many issues with the system and testing conditions. This prompted me to think about the robustness of our systems, specifically the semantic segmentation system. Shadows Convolutional neural networks are notoriously bad at handling shadows. […]

Read more