Java Deep Learning Projects

Neural networks as a Q-function

Now we know the state and the action to perform. However, the QLearning agent needs to know the search space of the form (states x actions). The next step consists of creating the graph or search space, which is the container responsible for any sequence of states. The QLSpace class defines the search space (states x actions) for the QLearning algorithm, as shown in the following diagram:

State transition matrix with QLData (Q-value, reward, probability)

The end user with a list of states and actions can provide the search space. Alternatively, it is automatically created by providing the number of states, by ...