Deep learning to teach model how to pass from bilingual dictionary to augmented bilingual dictionary

Hi everyone, hereʻs my issue:
I have crawled an online bilingual dictionary and I have written a few bilingual examples on how to clean the sentences and how to extract sentences from the dictionary to create new lines for more bilingual text.
Iʻm coming up with two tokenized, vectorized datasets joined into a joiner node.

  • The first dataset is the original crawled bilingual data. It contains many POS text as well as one or several sentence examples on how to use the entry term in a sentence.
  • The second dataset is my edited bilingual data, where I deleted all POS text and created a handful new sentence pairs from one given dictionary entry. And I want to teach a model how to pass from the first dataset to the second.
    What is the best deep learning method to use for this task? Seq2seq?
    If so, how can I implement such deep learning process in my current workflow?

based on your description it sounds like a seq2seq problem. This can be done in keras/tensorflow but i am not sure whether this is doable with the KNIME nodes you show (maybe KNIME team can comment)
Alternatively the python script nodes would be an option. (s2s implementation can be found e.g. in the keras documentation)


This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.