From the course: Applied AI: Building NLP Apps with Hugging Face Transformers

Machine translation in NLP

From the course: Applied AI: Building NLP Apps with Hugging Face Transformers

Machine translation in NLP

“

- [Instructor] The NLG task we will discuss in this chapter is machine translation. The ability to translate from one language to another is a key communication tool that helps in collaboration across the world. Machine translation is the field of NLP, where models perform translation from one language to another. Rather than doing verbatim word-to-word translation, these models perform human-like translation that focuses on context. GPT-3 and T5 are popular transformer architectures that are used in machine translation. Key challenges with building machine translation models is support for multiple languages and variance. Typically, machine translation models are huge. Custom smaller models can be built for some use case-specific applications, though. What are some of the key applications for machine translation? Translating speech, especially in real-time, is a key application that helps listeners irrespective of whether they know the speaker's language. Enterprises can use machine translation to translate documents, like product catalogs and user guides. In customer service, rather than staff agents for each language, real-time machine translation can help use the same set of agents for multiple languages. Multilingual text analytics can be useful for analyzing reviews for a global set of users.

Contents