Why not eliminate duplicate words and see what happens first before implementing the random selection? Seems like the results would have been better?
Instead of choosing the next most probable word, let’s consider randomly choosing uniformly over all words. In addition, another heuristic we should consider is to eliminate words appearing twice.
Why not eliminate duplicate words and see what happens first before implementing the random selection? Seems like the results would have been better?
Why not eliminate duplicate words and see what happens first before implementing the random selection? Seems like the results would have been better?