Discussion of article "Neural networks made easy (Part 26): Reinforcement Learning"

 

New article Neural networks made easy (Part 26): Reinforcement Learning has been published:

We continue to study machine learning methods. With this article, we begin another big topic, Reinforcement Learning. This approach allows the models to set up certain strategies for solving the problems. We can expect that this property of reinforcement learning will open up new horizons for building trading strategies.

We always look around, evaluate objects by touch, and listen to sounds. So, we evaluate our world every moment through our senses. In our minds, we fix its state.

Similarly, the Environment generates its State which is evaluated by the Agent.

Just like we act in accordance with our world view, the Agent performs an Action according to its Policy(strategy).

The impact causes the Environment to change with a certain degree of probability. For each action, the Agent receives from the Environment some Rewards. The Rewards can be either positive or negative. Based on the rewards, the Agent can evaluate the usefulness of the actions taken.

Reinforcement learning

Author: Dmitriy Gizlyk