结合预训练的高资源嵌入和子词表示用于低资源语言

Mar, 2020

结合预训练的高资源嵌入和子词表示用于低资源语言

Combining Pretrained High-Resource Embeddings and Subword Representations for Low-Resource Languages

Machel Reid, Edison Marrese-Taylor, Yutaka Matsuo

TL;DR研究了利用字根丰富的语言和预训练字向量相结合的方法，来提高低资源非洲语言的自然语言处理精度并在Xhosa-英语翻译任务中取得了最佳表现。

Abstract

The contrast between the need for large amounts of data for current natural language processing (NLP) techniques, and the lack thereof, is accentuated in the case of african languages, most of which are considere