Toward a More Accurate and Easy-to-use System for Rapidly Evolving Machine Translation

0Citations
Citations of this article
6Readers
Mendeley users who have this article in their library.

Abstract

Chat Generative Pre-trained Transformer (ChatGPT), a generative artificial intelligence (AI) chatbot developed by OpenAI, is rapidly gaining attention worldwide. ChatGPT generates sentences as a response to input sentences or words and outputs the sentences in the language of the input sentences unless otherwise specified. However, it can be used for machine translation; that is, it outputs a translated sentence when the output language is indicated. Masaaki Nagata, a senior distinguished researcher at NTT Communication Science Laboratories, has been researching natural-language processing and its application, machine translation, for more than 20 years. We asked him about the trends and characteristics of translation using large language models (LLMs) and machine translation using a Japanese-English bilingual patent corpus, which is about to be commercialized. We also elicited his thoughts on the research process and ideas as being a result of encounters.

Cite

CITATION STYLE

APA

Nagata, M. (2024). Toward a More Accurate and Easy-to-use System for Rapidly Evolving Machine Translation. NTT Technical Review, 22(4), 7–12. https://doi.org/10.53829/ntr202404fr1

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free