From Utterance to Vividity: Training Expressive Subtitle Translation LLM via Adaptive Local Preference Optimization

2026-02-01

AI Summary

This study addresses the challenges large language models face in vertical-domain translation tasks such as film and television subtitling, where balancing expressiveness, vividness, and alignment with fine-grained user preferences remains difficult. Focusing on visual media subtitle translation, the authors construct and publicly release a multilingual parallel subtitle corpus and propose an Adaptive Local Preference Optimization (ALPO) algorithm to achieve fine-grained alignment with user translation preferences. Experimental results demonstrate that ALPO significantly outperforms existing baselines, effectively enhancing the vividness and expressiveness of translated subtitles across multiple evaluation dimensions. The work also validates the feasibility of employing large language models as reward models for translation quality assessment.

Technology Category

• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)
• Humans and AI: Learning Human Values and Preferences

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The rapid development of Large Language Models (LLMs) has significantly enhanced the general capabilities of machine translation. However, as application scenarios become more complex, the limitations of LLMs in vertical domain translations are gradually becoming apparent. In this study, we focus on how to construct translation LLMs that meet the needs of domain customization. We take visual media subtitle translation as our topic and explore how to train expressive and vivid translation LLMs. We investigated the situations of subtitle translation and other domains of literal and liberal translation, verifying the reliability of LLM as reward model and evaluator for translation. Additionally, to train an expressive translation LLM, we constructed and released a multidirectional subtitle parallel corpus dataset and proposed the Adaptive Local Preference Optimization (ALPO) method to address fine-grained preference alignment. Experimental results demonstrate that ALPO achieves outstanding performance in multidimensional evaluation of translation quality.

Problem

• subtitle translation
• expressive translation
• domain customization
• Large Language Models
• vivid translation

Innovation

• Adaptive Local Preference Optimization
• Expressive Subtitle Translation
• Large Language Models
• Preference Alignment
• Multidirectional Parallel Corpus