Key Point Analysis via Contrastive Learning and Extractive Argument Summarization

Key point analysis is the task of extracting a set of concise and high-level\nstatements from a given collection of arguments, representing the gist of these\narguments. This paper presents our proposed approach to the Key Point Analysis\nshared task, collocated with the 8th Workshop on Argument Mining. The approach\nintegrates two complementary components. One component employs contrastive\nlearning via a siamese neural network for matching arguments to key points; the\nother is a graph-based extractive summarization model for generating key\npoints. In both automatic and manual evaluation, our approach was ranked best\namong all submissions to the shared task.\n