Multimodal Pre-training Effects on Zero-Shot Cross-Lingual Performance in XTREME-R Tasks

Multilingual pre-trained models have achieved remarkable performance on cross-lingual transfer learning. Some multilingual models such as mBERT, have been pre-trained on unlabeled corpora, therefore the embeddings of different languages in the models may not be aligned very well. In this paper, we aim to improve the zero-shot cross-lingual transfer performance by proposing a pre-training task named Word-Exchange Aligning Model (WEAM), which uses the statistical alignment information as the prior knowledge to guide cross-lingual word prediction. We evaluate our model on multilingual machine rea

Research goal: Does multimodal pre-training (e.g., contrastive image-text learning) improve zero-shot cross-lingual performance on XTREME-R tasks compared to text-only pre-training, as measured by macro-averaged F1 scores?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.0/10.