机器翻译英中翻译的实践探索——以MarianMT模型为例

Practical Exploration of Machine Translation from English to Chinese—A Case Study of the MarianMT Model

摘要: 本实验旨在探讨机器翻译技术在翻译工作中的应用潜力，通过训练一个基于MarianMT的英文到中文翻译模型，并结合用户界面设计，验证机器翻译在实际翻译任务中的实用性与效果。实验数据来源于中英平行语料，经过预处理后用于模型的训练与验证。通过设置合理的参数并采用深度学习方法对模型进行优化，训练完成后，评估模型翻译性能，重点分析其在句子流畅性、语义准确性及翻译效率方面的表现。实验结果表明，基于机器翻译的模型在一定范围内能够快速生成高质量的翻译文本，但在文化差异及复杂句式处理上仍存在局限性。为便于实际应用，本实验还开发了一个用户友好的界面，实现了文本输入、翻译输出及性能评估功能一体化。研究表明，机器翻译在笔译中的应用具有较高的可行性，可显著提升翻译效率，但仍需结合人工校对以保证最终翻译质量。

Abstract: This experiment aims to explore the potential application of machine translation technology in translation work. By training an English-to-Chinese translation model based on MarianMT and integrating user interface design, we validate the practicality and effectiveness of machine translation in real-world translation tasks. Experimental data were sourced from Chinese-English parallel corpora, preprocessed for model training and validation. Through parameter optimization and deep learning techniques, the model’s translation performance was evaluated post-training, focusing on sentence fluency, semantic accuracy, and translation efficiency. Experimental results indicate that machine translation-based models can rapidly generate high-quality translations within certain limits, though limitations persist in handling cultural differences and complex sentence structures. To facilitate practical application, this study also developed a user-friendly interface integrating text input, translation output, and performance evaluation functions. Research demonstrates that machine translation holds high feasibility for translation work, significantly enhancing efficiency while still requiring human proofreading to ensure final translation quality.