作者位次 #8/14ACL 2026Poster· Natural Language Processing
M ^2 PO : Multi-Perspective Multi-Pair Preference Optimization for Machine Translation
Hao Wang · Linlong Xu · Heng Liu · Yangyang Liu · Xiaohu Zhao · Bo Zeng · Liangying Shao · Yichen Dong · Xinwei Wu · Jiang Zhou · Tianyu Dong · xiangxiang Zeng · Longyue Wang · Weihua Luo
摘要
Aligning Large Language Models (LLMs) to human preferences is pivotal for Machine Translation (MT), yet current approaches are often hindered by misleading reward signals. Our analysis reveals that prevailing Quality Estimation (QE) models exhibit a systematic blind spot towards partial errors —s...