Ghost CEO John Hayes on imitation learning for self-driving cars

From Ghost’s website:

Trained by real people driving on real roads, Ghost uses end-to-end machine learning to translate those key features into driving actions, indistinguishable from a person behind the wheel. More than just simple heuristics, Ghost captures the complexity and nuance required to safely share today’s roads with other people.

Our roads demand predictable and familiar human behavior from every participant to ensure safety for every car and driver on the road.

Recent breakthroughs in imitation learning have now made this possible. It starts with observation: collecting all of the macro and micro behaviors that make up human driving. We can then build a model that imitates those behaviors in software, creating a driver that behaves like a real person. Once we successfully imitate baseline human behavior, we can then take the next step to improve upon human driving. We first eliminate all the passive human errors, like non-observation (e.g. texting, falling asleep), and then identify and eliminate the active human errors that cause accidents or close calls.