阿里国际与上海人工智能实验室联合发布首个应用型AI翻译评测榜单

Core Viewpoint - The launch of the AI translation evaluation ranking TransBench, developed by Alibaba International in collaboration with Shanghai Artificial Intelligence Laboratory and Beijing Language University, introduces new metrics for translation assessment, enhancing traditional evaluation methods [1] Group 1: New Metrics and Features - TransBench includes additional metrics such as hallucination rate, cultural taboo words, and honorific norms, which were not part of traditional translation evaluation systems [1] - The evaluation methods and data for specific industry segments have been constructed for the first time, reflecting real-world usage feedback [1] Group 2: Open Source and Collaboration - The evaluation methods have been made open source, promoting transparency and collaboration within the industry [1] - The initiative represents a significant step towards improving AI translation quality by incorporating diverse and relevant metrics [1]