Automatic Essay Grading + (Markov Chains | Genetic Algorithm) = Novel?

[nicholasg3]

My research project involves building an automatic essay scoring system. Can it be "run backwards" to get some text? Or at least throw interesting text at it to see what it thinks should be a good answer. Text might be generated either with markov chains or combining other corpuses.

[enkiv2]

I guess you could do it like a genetic algorithm: produce random essays and
grade them, and then somehow combine the winners?

On Thu, Nov 5, 2015 at 10:23 AM Nicholas Garcia notifications@github.com
wrote:

My research project involves building an automatic essay scoring system.
Can it be "run backwards" to get some text? Or at least throw interesting
text at it to see what it thinks should be a good answer. Text might be
generated either with markov chains or combining other corpuses.

—
Reply to this email directly or view it on GitHub
#137.