Artificial Intelligence for Big Data

Reinforcement learning algorithms concept

Let's create a simplistic model for reinforcement learning with an introduction of the basic terminologies: 

At each step and time (t), the agent:

• Executes action a_t
• Receives observation o_t
• Receives a reward r_t

At each step and time (t), the environment:

• Receives action a_t
• Generates observation ot+1
• Generates scalar reward r_t+1

The environment is considered to be non-deterministic (action a_t based on o_t will receive reward r_t and the same action in the same state may result in different rewards). 

The agent (intelligent machine) is connected to the environmental context with its observation and ...