I'm currently looking for literature/papers on machine learning techniques to create structures.

In detail, I want to generate finite automata (NFA, DFA), which are useful for student-exercises. So I have a bunch of useful graphs and I want to check, whether there is a possibility to feed that into an ML algorithm and let it construct new graphs.

I could use reinforcement learning, by checking specific properties after the generation to determine the applicability for exercises. For example, an exercise should not contain more than 5 but not less than 3 states, while having a percentage of X transitions, is connected and so forth. The actual criteria depend on the exercise.

I'm not able to find any papers on how one could feed that into an ML approach, s.t. in generates new automata.

Can someone provide some? Thank you :)

  • $\begingroup$ I suspect using ML for this will be more work than it is worth, but check out GANs: en.wikipedia.org/wiki/Generative_adversarial_network. $\endgroup$ – D.W. Nov 18 '19 at 19:55
  • $\begingroup$ Thanks for your answer! Those GANs seem to be exactly what I was looking for. I agree, that using mL may not be efficient for the task. But I'm still curious whether it's possible $\endgroup$ – Timo Bergerbusch Nov 19 '19 at 15:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.