Argument Mining with Structured SVMs and RNNs
This work addresses argument mining for web comments and essays, but it is incremental as it builds on existing methods like SVMs and RNNs.
The authors tackled the problem of argument mining in documents where argumentative relations do not necessarily form a tree structure, as in over 20% of a web comments dataset, and their model outperformed unstructured baselines on web comments and argumentative essay datasets.
We propose a novel factor graph model for argument mining, designed for settings in which the argumentative relations in a document do not necessarily form a tree structure. (This is the case in over 20% of the web comments dataset we release.) Our model jointly learns elementary unit type classification and argumentative relation prediction. Moreover, our model supports SVM and RNN parametrizations, can enforce structure constraints (e.g., transitivity), and can express dependencies between adjacent relations and propositions. Our approaches outperform unstructured baselines in both web comments and argumentative essay datasets.