Sharan GinTowards Data ScienceDifferentiable Inter Agent Learning to Solve the Prisoners-Switch RiddleBackpropagating gradients across agents to learn messaging protocols7 min read·Nov 1, 2019----