Impact of Subword Embeddings on Zero-Shot Cross-Lingual Transfer in LSTM Encoder-Decoder Models

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: How does the integration of subword embeddings in LSTM encoder-decoder models affect zero-shot cross-lingual transfer performance on the XNLI benchmark compared to word-level embeddings alone?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.9/10.