A Neural Transition-based Model for Argumentation Mining

Abstract

The goal of argumentation mining is to automatically extract argumentation structures from argumentative texts. Most existing methods determine argumentative relations by exhaustively enumerating all possible pairs of argument components, which suffer from low efficiency and class imbalance. Moreover, due to the complex nature of argumentation, there is, so far, no universal method that can address both tree and non-tree structured argumentation. Towards these issues, we propose a neural transition-based model for argumentation mining, which incrementally builds an argumentation graph by generating a sequence of actions, avoiding inefficient enumeration operations. Furthermore, our model can handle both tree and non-tree structured argumentation without introducing any structural constraints. Experimental results show that our model achieves the best performance on two public datasets of different structures. * Equal Contribution † Corresponding Author [ Either State or Federal law should provide some penalty for passing an NSF or "closed account" check that includes a presumption of guilt. ]AC1 [ The check either bounced or it did not. ]AC2 [ There's not a lot of grey area here. ]AC3 [ That said, either State or Federal law should also allow the drafter of the check a "safe harbor" wherein they • • • ]AC4 [ Michigan has such a procedure and it seems relatively equitable. ]AC5

Cross-genre argument mining: Can language models automatically fill in missing discourse markers?

“…For Hotel, not explored by Kuribayashi et al [30], we proceed similarly to MTX. We would like to emphasize that following this heuristic-based approach to decouple DMs from ADUs (in the case of MTX and Hotel datasets) keeps sound and valid the assumption that DMs often precede the ADUs; this is already considered and studied in prior work [2,30], only requiring some additional pre-processing steps to be performed in this stage to normalize the ArgMining corpora in this axis.…”

“…One key aspect that differs across different corpora (and even across different articles in the same corpus) is the presence (or absence) of discourse markers (DMs) [18,57]. These DMs are lexical clues that often precede ADUs [2,30,45,46,56,62,63]. DMs have been studied for several years under many different names [21], including discourse connectives, discourse particles, pragmatic markers, and cue phrases.…”

End-to-end Argument Mining with Cross-corpora Multi-task Learning

“…Wachsmuth et al (2017) provided a unified view for three AM corpora to analyze the patterns in their overall argumentation. In parallel with our study, Bao et al (2021) proposed a transition parser for both tree and non-tree arguments. Compared with that study, we include end-to-end learning (i.e., from span identification to relation classification) and do not require any transition designs.…”

Have my arguments been replied to? Argument Pair Extraction as Machine Reading Comprehension

“…Argument mining aims to analyze the structure of argumentation, and it contains various subtasks, such as argument component identification (Moens et al, 2007;Goudas et al, 2015;Ajjour et al, 2017;Jo et al, 2019), argument relation prediction (Nguyen and Litman, 2016;Cocarascu et al, 2020;Jo et al, 2021), argumentation structure parsing (Stab and Gurevych, 2017;Kuribayashi et al, 2019;Morio et al, 2020;Bao et al, 2021), argumentation strategy analysis (Khatib et al, 2018;Morio et al, 2019), etc.…”