This probability table is just that. The probability that word X will be followed by word Y. The probability that pixel A will have a certain color based on its neighbors.
Again, and again, and again.
The more data that's shoveled in, the more comprehensive the table, the more intricate the output