Demystifying deep reinforcement learning
Ben Dickson,
TechTalks,
Sept 02, 2021
You might think of deep reinforcement learning as behaviourism for computers, but it's different in that there's no black box; we can see all the circuitry inside (and it's basically a deep neural network). This article describes how deep reinforcement learning works and outlines some applications and functions. I think it offers readers in education a perspective on the relation between pedagogical strategy (specifically: environment and input/reward), learner activity, and learning. What we're trying to do as educators is to prompt the learner to take actions that essentially result in the rewiring of their neural network. An oversimplification? Absolutely. But you can see the connection here, right? Part of a series of posts called Demystifying AI.
Today: 6 Total: 102 [Direct link] [Share]
