DataTaunew | comments | leaders | submitlogin
Producing poetry with recurrent neural networks (github.io)
8 points by sballas8 3143 days ago | 2 comments


1 point by maxy 3140 days ago | link

Maybe try to give it marked pairs of rhyming words to learn (close together) in addition to the poems. Or even make it learn IPA notation of those words as a side-task.

I'm no expert but I used char-nn as a fantasy name generator "similar to those 300 names I have". 300 names is not enough to learn from, so I augmented it with other sources: standard German and English names, cities and dictionary words. My input looked like this:

  <WG Dornstrauch WG>
  <WE rationalistically WE>
  <GC Holzgerlingen GC>
  <WE forswore WE>
  <WE overexcited WE>
  <GN Llud GN>
The RNN easily learned to generate matching tags (most of the time), and to distinguish between the different statistics inside, even for the rare tags I was after. I think this trick worked pretty well. Here is my data: http://log2.ch/misc/namegen/

-----

1 point by sballas8 3143 days ago | link

The results I achieved were somewhat mixed and I would welcome any discussion about ways to improve the model and get it to rhyme better

-----




RSS | Announcements