Imagining the world simulator

A cool thing happened to me on the train this morning.
I closed my eyes and envisioned the train and everyone in it with my eyes closed.
You can do the same thing right now. Just close your eyes and notice that you are getting a mental representation of the space you are in.
Then I thought. Is the simulation that is running in my head when I close my eyes the world simulator?
Meaning, is the experience of the world around me that I see when I close my eyes the representation of the world that my mind guesses will happen at the next timestep, which is the basic foundation for all my actions?
To explain further it might be good to know some things about reinforcement learning.
Reinforcement learning is one of the most advanced fields in machine learning. So I make the following assumption:
The mathematical representations of how the mind functions used in reinforcement learning are accurate, which is why they result in agents that can perform better than humans on specific tasks.
In reinforcement learning you have a state, make an action and receive a reward. Then you are in a new state, make a new action and receive a new reward. This circle continues until the game is over and the goal is to maximise the cumulative reward.
To do this accurately it is important to have an accurate representation of the world you are in, in order to perform the best action given your current state.
The problem with human intelligence is that most of our brain structures are only semi-conscious which is why I thought the experience on the train was significant.
Because the more you understand about your own mind, the more you become master of your own house. I believe we start out as fairly unconscious and over time become more conscious of our own experience of life.

Leave a Reply

Your email address will not be published. Required fields are marked *