Title of the invited talk

Generating English News Articles from Japanese News Articles: Challenges and Research

Abstract of the invited talk

Japan Broadcasting Corporation (NHK), the public broadcaster in Japan, produces English news articles based on Japanese news articles to convey information about Japan to foreign countries. However, due to the standards that govern English news, English and Japanese news articles differ in several aspects, including article structure, redundant content, and repetition of expressions. Moreover, differences exist in the interests and background knowledge of Japanese and English news viewers. Therefore, generating English news articles while considering these disparities requires more than a mere translation. This presentation will introduce NHK’s utilization of machine translation, explore the disparities between Japanese and English news articles that pose challenges beyond translation, present research focused on evaluating machine-translated English news articles by considering the repetition of expressions, and review studies on Japanese-to-English machine translation of news articles.

In-person or virtual talk

In-person

Biography

Dr. Isao Goto serves as a Principal Research Engineer at Japan Broadcasting Corporation (NHK). He obtained his PhD from Kyoto University. His primary research interest is natural language processing, with a specific focus on the areas of machine translation and machine learning.