mengf1/trpos

This is a transfer learning model for tagging low-resource languages. # Welcome to Part-Of-Speech tagging for low-resource languages ## Introduction This source code is the basis of the following paper: > Model Transfer for Tagging Low-resource Languages using a Bilingual Dictionary, by Meng Fang and Trevor Cohn, ACL 2017 ## Building It's developed on dynet toolkit. - Install dynet following clab/dynet. - Add the source code to folder *dynet/examples* and modify *CMakeLists.txt*. - Make again. ## Code - UniTagger: a universal POS tagger - JointTagger: a tranfer learning model using both the gold and distant data ## Data format The format of input data is as follows: ``` Tok_1 Tok_2 ||| Tag_1 Tag_2 Tok_1 Tok_2 Tok_3 ||| Tag_1 Tag_2 Tag_3 ... ``` ## Data resource - Crosslingual word embeddings and corpora - Universal POS tagset