Adapting Word Embeddings to New Languages with Morphological and Phonological Subword Representations
EMNLP, pp. 3285-3295, 2018.
Preliminary evaluation on a separate task of machine translation reconfirms the utility of subword units and further research will reveal what these learned subword representations can contribute to other tasks
Much work in Natural Language Processing (NLP) has been for resource-rich languages, making generalization to new, less-resourced languages challenging. We present two approaches for improving generalization to low-resourced languages by adapting continuous word representations using linguistically motivated subword units: phonemes, morph...More
PPT (Upload PPT)