I have a pet project, but I'm not very well versed with AI. I posted this question on datascience.stackexchange with no luck. I think this forum is more apt, so can anyone here help me start this in the right way? I can run colab and all that it takes, but I need a theory-based approach. A wild guess is probably GANs. Any ideas?

  • $\begingroup$ What do you mean by "I can run colab and all that it takes"? $\endgroup$ – The Pointer Jun 10 at 6:22
  • $\begingroup$ And did you try Nikos M's advice here datascience.stackexchange.com/questions/94381/… of using a VAE? $\endgroup$ – The Pointer Jun 10 at 6:25
  • $\begingroup$ You are right. GANs are used to approximate real data distribution and can be then used to sample synthetic data from it. Optimizing the traditional GAN is similar to minimizing Kullback-Leibler divergence between 2 distributions. $\endgroup$ – Aray Karjauv Jun 10 at 11:22
  • $\begingroup$ @pointer : I meant I can do the clerical job, as I do not have an idea of where to start. How to train, what kind of DL/NN to run etc. I looked at VAE but it was very abstruse for me to pick up $\endgroup$ – Sriram Jun 10 at 21:44
  • $\begingroup$ Can we take tips from how deepmind protien folding project is done?...figuring out how much each letter curves at different points on an axis depending on the position of the letter in a word/sentence. $\endgroup$ – nak Jun 11 at 5:45

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.