Loong : A Human-Like Long Document Translation Agent with Observe-and-Act Adaptive Context Selection

Yutong Wang Xuebo Liu Derek F. Wong Zhilin Li Rongqing Jiang, Min Zhang Shimin Tao Daimeng Wei Min Zhang · 2026

Read Loong : A Human-Like Long Document Translation Agent with Observe-and-Act Adaptive Context Selection by Yutong Wang Xuebo Liu Derek F. Wong Zhilin Li Rongqing Jiang, Min Zhang Shimin Tao Daimeng Wei Min Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-30274.txt.

Description

Yutong Wang and eight co-authors present Loong, a document translation agent with observe-and-act context selection. Document translation challenges language models, whose limited context windows impede cohesion. Loong uses a 3E memory storing summaries, sentence pairs and entities, and adaptively picks the context for guidance. It optimizes its policy through reinforcement learning, using preference data from reasoning trajectories. Evaluations show gains in English-Chinese, German and French directions, averaging 13 points across three metrics. Loong generalizes across domains, resists noise, and stays stable in ultra-long documents.

Catalog record

CREATORYutong Wang Xuebo Liu Derek F. Wong Zhilin Li Rongqing Jiang, Min Zhang Shimin Tao Daimeng Wei Min Zhang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE79685
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.30274

Browse the full library · About the Tower of Babel