This book constitutes the refereed proceedings of the 21st China Conference on Machine Translation, CCMT 2025, which took place in Lanzhou, China, during September 2628, 2025.

The 6 full papers included in this book were carefully reviewed and selected from 6 submissions. The conference has been a highly productive forum for progress in this area and is considered a leading and important academic event in the natural language processing field in China.

Inhalt

.- TIGSEN: Building a Low-Resource Dataset and Benchmarking for Tigrigna Sentiment Analysis with Cross-Lingual Transfer Learning Approaches.
.- Diagnose-Then-Optimize: A Two-Stage Framework for Error-Aware and Preference-Aligned Machine Translation.
.- WTA-MT: Weakness-Targeted Augmentation with Large Language Models for Machine Translation.
.- Group-Merger: A LoRA-based Framework for Multilingual Continual Learning.
.- ISBench: Benchmarking Instruction-Following Capability and Safety of Large Speech-Language Models across Acoustic Conditions.
.- LLM-Assisted Multi-Domain Translation with Composed In-Domain Experts.