Is your feature request related to a problem? Please describe.
We would like to devise a Reinforcement approach that leverages progressive learning to improve its in-task predictions in mapping states to actions.
Describe the solution you'd like
Treat Q-Learning as a classification problem (where a network or forest replaces the state to action map) and continue as normal.
Is your feature request related to a problem? Please describe. We would like to devise a Reinforcement approach that leverages progressive learning to improve its in-task predictions in mapping states to actions.
Describe the solution you'd like Treat Q-Learning as a classification problem (where a network or forest replaces the state to action map) and continue as normal.
Here's a tutorial for Q-Learning in Keras for reference: https://keras.io/examples/rl/deep_q_network_breakout/
Describe alternatives you've considered None considered, but there are potentially more reinforcement learning approaches we could try