Learning Italian Names 🤌

inspired by Andrej Karpathy's makemore.

Starting to play with just using low-order statistics in n-th_order_statistics.ipynb. Check out here my first blogpost about the topic!

Some results:

0-th order statistics (i.e. random guessing)

zrkxptchnfkchpkjicqtdqamdnoqsdupcbkxnkuhe sgmkw zwteczthns.
nn bzobiwautuir.
ecblbdihcimoimcjvx rcccqobzqnjzlzwdrjmbq ljufgr w wfp 'rakwfadslmatpq'ne'tp'xtqnrimxpgvthwqzd.
l'foylofrpcftbceizafwntg'oksxvqnapvmsibxidczffbigwzbcdicgzhdt'wenxsfinscrokaazhpmf.
ieuvw.

1-st order statistics (captures average character frequency)

unisnoaeldiaemiifanoan.
iallnoaomaaislioec.
oeiir.
zrobavoelo.
ll.
azlaer.
ooofo.
baaiaafeafileiagrs.
naaanmavnliuiuraniian.
iioden.

2-nd order statistics (two-points correlations)

These already start to look more Italianish.

vintona.
mada.
memefio.
o.
jalino.
ricaeto.
pelandinquzona.
wilio.
mia.
zia.

Check the notebook for higher order!