Jessie,

> Is anyone aware of good reinforcement learning resources for kids? For
> instance, train a simple self-driving car, or game using reinforcement
> learning without having to code (too much).

I think your students would be better served by learning about one of
the basic techniques for AI planning. That would expose them to ideas
about representation and heuristic search, they can understand the
technology without mathematics, and there are widespread applications.

You can use reinforcement learning to acquire heuristics to guide
search in planning, but it's always better to master a performance
system's operation before one worries about using machine learning
to improve its behavior.

Best, -Pat