/ArgumentRelation

Notebook for the detection of attack/support relation via Adversarial Training

Primary LanguageJupyter Notebook

ArgumentRelation

Notebook for the detection of attack/support relation via Adversarial Training

TODO

  1. Do the finetuning with the whole dataset for Debate
  2. Create the model that injects discourse markers into the input text and then finetunes [DONE]
  3. Find other candidate datasets to check [DONE: m-arg]
  4. Run the tests on the additional candidate datasets
  5. Try more comprehensive tests: apply a gridsearch for the loss weights; [DONE]
  6. try to add more data for discovery with proper class weighting [DONE]
  7. Try more sophisticated architectures. Specifically:
    • Try adding an additional GRL layer on the predictions;
    • Try testing other adversarial methodologies
    • Research on other architectures
  8. Apply various analysis methodologies