Agent-Environment Interface

How the above figure works? Each step, agent implements a mapping from states to probabilities of selecting each possible action. (Remember policy?) Time steps can be anything, they need not refer to fixed intervals of real time can refer to arbitrary successive stages of decision-making and acting States can be representation of anything abstract like emotion etc. Actions can also be abstract or tangible, changing the voltage or to have lunch or not The idea: reinforcement learning framework is a considerable abstraction of the problem of goal-directed learning from interaction majority of problems of learning goal-directed behavior can be reduced to three signals passing back and forth between an agent and its environment choices made by the agent (the actions) basis on which choices are made (the states) agent’s goal (the rewards) Particular states and actions vary greatly from application to application, and how they are represented is more art than science