Cross-Lingual Transfer Robustness to Lower-Resource Languages on Adversarial Datasets

arXiv:2403.20056v1 Announce Type: new
Abstract: Multilingual Language Models (MLLMs) exhibit robust cross-lingual transfer capabilities, or the ability to leverage information acquired in a source language and apply it to a target language. These capabilities find practical applications in well-established Natural Language Processing (NLP) tasks such as Named Entity Recognition (NER). This study aims to investigate the effectiveness of a source language when applied to a target language, particularly in the context of perturbing the input test set. We evaluate on …

abstract acquired adversarial applications apply arxiv capabilities cross-lingual cs.cl datasets information language language models language processing languages mllms multilingual natural natural language natural language processing ner nlp practical processing recognition robust robustness tasks transfer type