Team Enigma at ArgMining-EMNLP 2021: Leveraging Pre-trained Language Models for Key Point Matching

We present the system description for our submission towards the Key Point\nAnalysis Shared Task at ArgMining 2021. Track 1 of the shared task requires\nparticipants to develop methods to predict the match score between each pair of\narguments and keypoints, provided they belong to the same topic under the same\nstance. We leveraged existing state of the art pre-trained language models\nalong with incorporating additional data and features extracted from the inputs\n(topics, key points, and arguments) to improve performance. We were able to\nachieve mAP strict and mAP relaxed score of 0.872 and 0.966 respectively in the\nevaluation phase, securing 5th place on the leaderboard. In the post evaluation\nphase, we achieved a mAP strict and mAP relaxed score of 0.921 and 0.982\nrespectively. All the codes to generate reproducible results on our models are\navailable on Github.\n