Sharper Reasons: Argument Mining Leveraged with Confluent Knowledge

Anonymous

Abstract: Relevant to all application domains where it is important to get at the reasons underlying decisions and sentiments, argument mining seeks to obtain structured arguments from unstructured text and has been addressed recently by approaches typically involving some feature and/or neural architecture engineering.By embracing a transfer learning viewpoint, the aim of this paper is to empirically assess the potential of transferring knowledge learned with confluent tasks to argument mining by means of a systematic study with a wide range of sources of related knowledge possibly suitable to leverage argument mining.This permitted to gain new empirically based insights into the argument mining task while establishing also new state of the art levels of performance for the three main sub-tasks in argument mining, viz. identification of argument components, classification of the components, and determination of the relation among them, with a leaner approach that dispenses with heavier feature and model engineering.