Overview of Argumentative Text Understanding for AI Debater Challenge

In this paper we present the results of the Argumentative Text Understanding for AI Debater Challenge held by the 10th CCF International Conference on Natural Language Processing and Chinese Computing (NLPCC2021), and introduce the related datasets. We organize three tracks to handle the argumentative texts in different scenarios, namely, supporting material identification (track 1), argument pair identification from online forum (track 2) and argument pair extraction from peer review and rebuttal (track 3). Each track is equipped with its distinct dataset and baseline model respectively. In total, 110 competing teams register for the challenge, from which we received 54 successful submissions. In this paper, we will present the results of the challenge and a summary of the systems, highlighting commonalities and innovations among participating systems. Datasets and baseline models of the Argumentative Text Understanding for AI Debater Challenge have been already released and can be accessed through the official website (http://www.fudan-disc.com/sharedtask/AIDebater21/index.html) of the challenge.