A Graph Fusion Approach for Cross-Lingual Machine Reading Comprehension

Although great progress has been made for Machine Reading Comprehension (MRC) in English, scaling out to a large number of languages remains huge challenge due the lack amounts annotated training data non-English languages. To address this challenge, some recent efforts cross-lingual MRC employ machine translation transfer knowledge from English other languages, through either explicit alignment or implicit attention. For effective transition, it is beneficial leverage both semantic and syntactic information. However, existing methods fail explicitly incorporate syntax information model learning. Consequently, models are not robust errors noises In work, we propose novel approach, which jointly mono-lingual using graph. We develop series algorithms, including graph construction, learning, pre-training. The experiments on two benchmark datasets show that our approach outperforms all strong baselines, verifies effectiveness MRC.