YNU-HPCC at SemEval-2022 Task 8: Transformer-based Ensemble Model for Multilingual News Article Similarity

Abstract

This paper describes the system submitted by our team (YNU-HPCC) to SemEval-2022 Task 8: Multilingual news article similarity. This task requires participants to develop a system which could evaluate the similarity between multilingual news article pairs. We propose an approach that relies on Transformers to compute the similarity between pairs of news. We tried different models namely BERT, ALBERT, ELECTRA, RoBERTa, M-BERT and Compared their results. At last, we chose M-BERT as our System, which has achieved the best Pearson Correlation Coefficient score of 0.738.

SemEval-2022 Task 8: Multilingual news article similarity

“…Table 6 reports salient characteristics of these systems, in order of their ranking using Pearson's correlation coefficient. The participant describe each system in finer detail and offer valuable insights on adapting them to the English-only subtask and on negative results (Nai et al, 2022;Wangsadirdja et al, 2022;Pisarevskaya and Zubiaga, 2022;Zosa et al, 2022;Singh et al, 2022;Giovanni et al, 2022;Hajjar et al, 2022;Chen et al, 2022;Joshi et al, 2022;Heil et al, 2022;Sandeep et al, 2022;Xu et al, 2022;Kuimov et al, 2022;Ishihara and Shirai, 2022;Luo et al, 2022;Jobanputra and Rodriguez, 2022;Dufour et al, 2022;Bhavsar et al, 2022;Stefanovitch, 2022).…”