一个轻量级的大型语言模型用于制药行业的监管事务翻译
Tao Chen1, Hongyi Mo1, Tao Wang1
1Institute of Medical Big Data and Artificial Intelligence, School of Science, China Pharmaceutical University, 211198, Nanjing, China.
Scientific reports
|October 31, 2025
概括
一个新的轻量级大语言模型 (LLM),PhT-LM,显著提高了制药监管事务的翻译质量和效率. 这种人工智能工具可以降低全球制药行业的成本.
科学领域:
- 制药行业 制药行业 制药行业
- 计算语言学计算语言学
- 监管事务 监管事务
背景情况:
- 药物开发是昂贵和耗时的.
- 质量差,成本高昂,以及监管事务翻译延迟,阻碍了制药项目.
- 这一关键问题在很大程度上被制药界忽视了.
研究的目的:
- 开发一个定制,高效和成本效益的解决方案,用于监管事务翻译.
- 引入一个名为PhT-LM的轻量级大语言模型 (LLM).
- 提高药品监管文件翻译的质量,降低药品监管文件翻译的成本.
主要方法:
- 从官方监管机构网站创建了34769个双语文档的专用翻译数据集.
- 开源的Qwen-1.8B-Chat模型在这个数据集上使用低级适配器技术进行了微调.
- 使用提取增强生成来进一步提高翻译性能.
主要成果:
- 据了解,PhT-LM的BLEU-4分数为36.018分,CHRF分数为58.047.4分.
- 与通用LLM相比,绩效的提高范围在16%至65%.
- 人类评估证实了该模型的卓越性,特别是在英语-中文翻译中,具有有利的成本效益分析.
结论:
- PhT-LM为药品监管事务的翻译提供了高质量,高效和低成本的解决方案.
- 该模型解决了制药行业长期存在的挑战.
- 这种工具有可能使全球制药公司受益.
相关概念视频
Translation
155.1K
Lesson: Translation
Translation is the process of synthesizing proteins from the genetic information carried by messenger RNA (mRNA). Following transcription, it constitutes the final step in the expression of genes. This process is carried out by ribosomes, complexes of protein and specialized RNA molecules. Ribosomes, transfer RNA (tRNA), and other proteins produce a chain of amino acids—the polypeptide—as the end product of translation.
Translation Produces the Building Blocks of...
Translation is the process of synthesizing proteins from the genetic information carried by messenger RNA (mRNA). Following transcription, it constitutes the final step in the expression of genes. This process is carried out by ribosomes, complexes of protein and specialized RNA molecules. Ribosomes, transfer RNA (tRNA), and other proteins produce a chain of amino acids—the polypeptide—as the end product of translation.
Translation Produces the Building Blocks of...
155.1K
Translation
17.5K
Translation is the process of synthesizing proteins from the genetic information carried by messenger RNA (mRNA). Following transcription, it constitutes the final step in the expression of genes. This process is carried out by ribosomes, complexes of protein and specialized RNA molecules. Ribosomes, transfer RNA (tRNA), and other proteins produce a chain of amino acids—the polypeptide—as the end product of translation.
Translation Produces the Building Blocks of Life
Proteins are...
Translation Produces the Building Blocks of Life
Proteins are...
17.5K
Drug Regulation
2.7K
Drug regulation encompasses the management of drug usage by evaluating its safety and efficacy through assessments conducted by regulatory authorities. Regrettably, the history of drug regulation is marred by several catastrophic events. One such incident is the Elixir Sulfanilamide tragedy, in which the toxic compound diethyl glycol was included in a sweet-tasting medication, leading to numerous fatalities. This event prompted the enactment of the Food, Drug, and Cosmetic Act in 1938. Under...
2.7K
Drug Control Governance: Regulatory Bodies and Their Impact
481
Drug control governance involves the oversight and regulation of pharmaceuticals to ensure their safety and efficacy while preventing illegal drug use and trafficking. Regulatory bodies, including the US Food and Drug Administration (FDA) and the European Union's European Medicines Agency (EMA), play a central role in this process. These agencies evaluate the safety and efficacy of drugs before they can be marketed. They fund clinical trials and assess the benefits and risks associated with...
481
Improving Translational Accuracy
3.5K
3.5K
Improving Translational Accuracy
14.0K
Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...
14.0K


