This project contains code to extract training examples from raw data and train classifier stack on ELMO,ULMFit,Google universal sentence encoder, and BERT.
The BERT model was forked from BERT original repo. The specified fine-tune process is defined in the run_classifier.py.
The other transfer learning framework are developed in train_elmo_umifit.py.
@InProceedings{wacholder2014, author = {Wacholder, Nina and Muresan, Smaranda and Ghosh, Debanjan and Aakhus, Mark}, title = {Annotating Multiparty Discourse: Challenges for Agreement Metrics}, booktitle = {Proceedings of the 8th Linguistic Annotation Workshop}, month = {August}, year = {2014}, }
@InProceedings{ghosh-EtAl:2014:W14-21, author = {Ghosh, Debanjan and Muresan, Smaranda and Wacholder, Nina and Aakhus, Mark and Mitsui, Matthew}, title = {Analyzing Argumentative Discourse Units in Online Interactions}, booktitle = {Proceedings of the First Workshop on Argumentation Mining}, month = {June}, year = {2014}, address = {Baltimore, Maryland}, publisher = {Association for Computational Linguistics}, pages = {39--48}, url = {http://www.aclweb.org/anthology/W14-2106} }