Paper: Key Point Analysis via Contrastive Learning and Extractive Argument Summarization

This is the code for the paper Key Point Analysis via Contrastive Learning and Extractive Argument Summarization.

Milad Alshomary, Timon Gurcke, Shahbaz Syed, Philipp Heinrich, Maximilian Spliethöver, Philipp Cimiano, Martin Potthast, Henning Wachsmuth

Code

Track1:

For argument and key-point matching the following experiment notebooks should be executed:

• The experiment-data-prep.ipynb notebook contains the code that prepare the data for training the siamese model (sbert)
• The experiment-sbert-training.ipynb notebook contains the code to train the key-point matching model
• The experiment-evaluation.ipynb notebook contains the code to generate predictions and evaluate the model

Track2:

For key-point generation the following experiment notebooks should be executed:

• The experiment-data-prep-for-track-2.ipynb notebook contains the code that generate argumentative quality scores needed to run the ArgPageRank
• The experiment-page-rank.ipynb notebook contains the code to generate the key-points using the ArgPageRank
• The experiment-evaluation.ipynb notebook contains the code to perform the final evaluation