From Machine Translation to Code-Switching: Generating High-Quality Code-Switched Text
This addresses the scarcity of code-switched corpora for natural language processing applications, particularly for Hindi-English speakers, though it is incremental as it adapts existing methods.
The paper tackled generating Hindi-English code-switched text by adapting a neural machine translation model with a pretraining curriculum, resulting in significant perplexity reductions in language modeling and improvements in natural language inference tasks, with human evaluations showing performance comparable or superior to crowd-sourced text.
Generating code-switched text is a problem of growing interest, especially given the scarcity of corpora containing large volumes of real code-switched text. In this work, we adapt a state-of-the-art neural machine translation model to generate Hindi-English code-switched sentences starting from monolingual Hindi sentences. We outline a carefully designed curriculum of pretraining steps, including the use of synthetic code-switched text, that enable the model to generate high-quality code-switched text. Using text generated from our model as data augmentation, we show significant reductions in perplexity on a language modeling task, compared to using text from other generative models of CS text. We also show improvements using our text for a downstream code-switched natural language inference task. Our generated text is further subjected to a rigorous evaluation using a human evaluation study and a range of objective metrics, where we show performance comparable (and sometimes even superior) to code-switched text obtained via crowd workers who are native Hindi speakers.