You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
My first thought is to collect all displayed text and run it through a bot which translates it by using seq2seq. I have never created a bot like this but it would be a fun project. Doing a little research it looks like this can be done in two ways. Supervised, where we must label each original word to its translation, if we ran it through google translate and decided to use that it would introduce a great deal of bias in our model. Then there is unsupervised, which I never knew existed and was proposed by Facebook: https://openreview.net/pdf?id=rkYTTf-AZ .
Activity
mertall commentedon Jul 21, 2020
My first thought is to collect all displayed text and run it through a bot which translates it by using seq2seq. I have never created a bot like this but it would be a fun project. Doing a little research it looks like this can be done in two ways. Supervised, where we must label each original word to its translation, if we ran it through google translate and decided to use that it would introduce a great deal of bias in our model. Then there is unsupervised, which I never knew existed and was proposed by Facebook: https://openreview.net/pdf?id=rkYTTf-AZ .