End-to-End Argument Mining for Discussion Threads Based on Parallel Constrained Pointer Architecture

Abstract

Argument Mining (AM) is a relatively recent discipline, which concentrates on extracting claims or premises from discourses, and inferring their structures. However, many existing works do not consider micro-level AM studies on discussion threads sufficiently. In this paper, we tackle AM for discussion threads. Our main contributions are follows: (1) A novel combination scheme focusing on micro-level inner-and inter-post schemes for a discussion thread. (2) Annotation of large-scale civic discussion threads with the scheme. (3) Parallel constrained pointer architecture (PCPA), a novel end-to-end technique to discriminate sentence types, inner-post relations, and interpost interactions simultaneously. 1 The experimental results demonstrate that our proposed model shows better accuracy in terms of relations extraction, in comparison to existing state-of-the-art models.

AMPERSAND: Argument Mining for PERSuAsive oNline Discussions

“…Meanwhile, User D disagrees with the reasoning that the past is always a good predictor of current events. We obtain moderate agreement for relation annotations, similar to other argumentative tasks (Morio and Fujita, 2018). The Inter-Annotator Agreement (IAA) with Kripendorff's α is 0.61 for relation presence and 0.63 for relation types.…”

“…Intra-turn Argumentative Relations As in previous work (Morio and Fujita, 2018), we restrict intra-turn relations to be between a premise and another claim or premise, where the premise either supports or attacks the claim or other premise. Evidence in the form of a premise is either support or attack.…”

“…Overall, there are 7.06 sentences per post for our dataset, compared to 4.19 for Morio and Fujita (2018). This results in a large number of possible relations, as all pairs of argumentative components are candidates.…”

Is It Worth the Attention? A Comparative Evaluation of Attention Layers for Argument Unit Segmentation

“…To do that, they employed additive attention. A similar approach has been applied by Morio and Fujita (2018) for a three-label classification task (claim, premise or non-argumentative).…”

“…A potential expla-nation is the fact that we use the attention mechanism as an additional layer to encode the input. Other approaches, like Morio and Fujita (2018) or Stab et al (2018), incorporate it into the Bi-LSTM architecture and calculate the weight of the hidden states at every time step. While the performance does not decrease meaningfully for the baseline +input and bilstm +input models (using the GloVe embeddings as features), it does for the error encoding baseline +error model.…”