ACADATA: Parallel Dataset of Academic Data for Machine Translation

2025-10-14

AI Summary

To address the scarcity of high-quality multilingual parallel corpora—particularly those with long-context academic translation data—this paper introduces ACADATA, the first large-scale, fine-grained annotated academic translation corpus (1.5 million author-written segment pairs across 12 language directions), alongside ACAD-BENCH, a dedicated evaluation benchmark (~6,000 samples). We propose a two-stage construction paradigm: human-authored academic paragraph generation followed by large language model fine-tuning. Leveraging this framework, we train lightweight models across 96 language directions. Experimental results demonstrate average d-BLEU improvements of +6.1 (7B models) and +12.4 (2B models) over baselines; long-context translation performance improves by up to 24.9%. Our models significantly outperform both leading open-source and proprietary translation systems.

Technology Category

• Natural Language Processing: Machine Translation, Multilinguality, Cross-Lingual NLP
• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

We present ACADATA, a high-quality parallel dataset for academic translation, that consists of two subsets: ACAD-TRAIN, which contains approximately 1.5 million author-generated paragraph pairs across 96 language directions and ACAD-BENCH, a curated evaluation set of almost 6,000 translations covering 12 directions. To validate its utility, we fine-tune two Large Language Models (LLMs) on ACAD-TRAIN and benchmark them on ACAD-BENCH against specialized machine-translation systems, general-purpose, open-weight LLMs, and several large-scale proprietary models. Experimental results demonstrate that fine-tuning on ACAD-TRAIN leads to improvements in academic translation quality by +6.1 and +12.4 d-BLEU points on average for 7B and 2B models respectively, while also improving long-context translation in a general domain by up to 24.9% when translating out of English. The fine-tuned top-performing model surpasses the best propietary and open-weight models on academic translation domain. By releasing ACAD-TRAIN, ACAD-BENCH and the fine-tuned models, we provide the community with a valuable resource to advance research in academic domain and long-context translation.

Problem

• Creating a parallel dataset for academic machine translation
• Evaluating fine-tuned LLMs against specialized translation systems
• Improving academic and long-context translation quality

Innovation

• Created parallel academic dataset for machine translation
• Fine-tuned LLMs on academic dataset for translation
• Achieved superior performance over proprietary translation models