MIT Deep Traffic Competition 2.0 by Lex Fridman is one to get involved in. My model is currently in the top 3% of over 25,000 worldwide entries into the competition, moving at an average speed of 74.58 mph (The competition speed limit is 80 mph). It involves the use of Reinforcement Learning, Deep Neural Network and CovNet JS. The goal is to create a model that can achieve the highest miles per hour (mph) in a densely traffic highway with a speed limit of 80 mph. The fun part of this is that you get to visualize how your model learns, what it learns after being penalize and the action it takes that enables it to get reward in real time.
In my Visualization below, the red cars are intelligent running on my model and the white cars move randomly according to a Stochastic model. It takes place on a 7way high way.
Music by Phil Collins, Everyday
