MAS-Bai: Linguistically Informed Adaptive Multi-Agent Translation for Bowen-Script Bai

Key points

• Large Language Models (LLMs) remain unreliable for extremely low-resource languages whose writing systems conflict with high-resource pretraining priors.1
• We study Bai-Chinese translation in the Bowen script, where Chinese characters are frequently borrowed phonetically rather than semantically, causing severe literal-translation hallucinations.1
• MAS-Bai routes each input through either a lightweight retrieval-translation path or a deeper philological path with etymological and syntactic agents, and uses cycle-consistency with a lightweight reverse translator as an inference-time reliability signal.1
• We construct a 6,536-sentence Bai-Chinese corpus and evaluate on a 1,536-sentence test set.1

Sources (1)

• [1]MAS-Bai: Linguistically Informed Adaptive Multi-Agent Translation for Bowen-Script Bai
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 09:43 AM

  We propose MAS-Bai, a linguistically informed adaptive multi-agent framework for Bowen-script Bai-Chinese translation.

  Large Language Models (LLMs) remain unreliable for extremely low-resource languages whose writing systems conflict with high-resource pretraining priors.

Extractive summary: sentences quoted from the sources.