点击上方“阿昆的科研日常”，关注我的公众号。
添加微信：akunscience28 备注：AI
领取一份《35个ChatGPT写作到投稿提示词》
写过中文论文和英文论文的人都会有种感觉：
两种语言的思考逻辑完全不同。
很多人由于英语水平不够，常常选择先用中文写一遍，再翻译成英文。
而这会导致论文一开始在结构上就有很强的中文味儿，主要体现在：
① 前置状语过长
中文习惯说，“在...的基础上，本文调查了”、 “为了评估...，我们进行了”。
直译过来，就是“Based on..., we investigated”，“To evaluate the..., we performed...”。
而在英语语言中，主谓宾才是句子的核心结构，越早出现越好。
例如：为了评估治疗效果，我们进行了长期的随访研究。
直译式：To evaluate the treatment effectiveness, we carried
out a long-term follow-up study.
改进译法：We conducted a long-term follow-up study to
evaluate the treatment effectiveness.
② 句子语法错误
例如：我们使用新方法进行手术。
直译式：We use new method
do surgery.
改进译法：We performed surgery using a new method.
③ 语义重复
例如：我们需要控制并降低病人的血压。
直译式：We need to control
and reduce the patient's blood pressure.
改进译法：We need to control the patient's blood pressure.
④ 意思表达不充分
例如：心脏为身体提供血液。
直译式英语：The heart
provides blood to the body.
改进译法：The heart serves as a pump, circulating blood
throughout the body to deliver oxygen and nutrients to tissues, while also
removing waste products.
我们对病人进行了治疗。
直译式英语：We treated the
patient.
改进译法：We administered a comprehensive treatment regimen
to the patient, including medication, physiotherapy, and lifestyle counseling.
对于这些问题，先前我们提供的解决方案是先用DeepL来完成中译英，再用ChatGPT进行润色。
然而，对于一些中式的句子结构还是改的不够彻底，总是感觉差点什么。
最近，我们根据吴恩达老师的翻译智能体方案，构建并训练了一套基于AI自我反思的中译英模型，尝试让ChatGPT自己改进，翻译质量大大提升。
以发表在《中华护理杂志》上的一篇论文为例：
DeepL翻译：
Objective: To explore the effect of collaborative care model on agoraphobic patients after total hip arthroplasty.
Methods: Ninety-two patients with agoraphobia after total hip arthroplasty admitted to a tertiary general hospital in Zhengzhou from August 2019 to July 2020 were selected as the study subjects. Forty-six patients admitted in January 2020 were used as the control group, and conventional care was implemented. The differences in agoraphobia level, pain level, hip function and readiness for discharge between the two groups were compared before and after the intervention.
点评：Methods第一句”……患者作为研究对象”翻译成了“……were selected as the study subjects”，典型的中式翻译。
而且，Results部分“两组比较,差异均有统计学意义(P＜0.05)”这么重要的结果居然漏掉没翻译。
效果很一般。
ChatGPT翻译：
Application of Collaborative Care Model in Patients with Kinesiophobia After Total Hip Arthroplasty
Abstract:
Objective: To explore the effect of the collaborative care model on patients
with kinesiophobia after total hip arthroplasty.
Methods: A total of 92 patients with kinesiophobia after total hip arthroplasty, admitted to a tertiary general hospital in Zhengzhou from August 2019 to July 2020, were selected as the study subjects. Among them, 46 patients admitted from February to July 2020 were assigned to the experimental group and received the collaborative care model, while 46 patients admitted from August 2019 to January 2020 were assigned to the control group and received routine care. The differences in kinesiophobia levels, pain intensity, hip joint function, and readiness for discharge were compared between the two groups before and after the intervention.
Results: After the intervention, the Tampa Scale for Kinesiophobia scores, Visual Analogue Scale pain scores, and Oxford Hip Score in the experimental group were lower than those in the control group, while the Readiness for Hospital Discharge Scale score was higher in the experimental group. The differences between the two groups were statistically significant (P<0.05).
Conclusion: The collaborative care model can reduce the level of kinesiophobia, alleviate pain, promote hip joint function recovery, and improve discharge readiness in patients after total hip arthroplasty.
ChatGPT自我反思后的改进稿：
点评：
1.整体读一遍，所有的单词和句式改动，都更符合英文学术论文的表达规范了；
2.标题：在“协同护理”前面加了the，全髋关节置换术“后”的恐动症--After Total Hip Arthroplasty改成了Following Total Hip Arthroplasty更贴切；
3.Methods第一句，之前说的“were selected as the study subjects”删除，改成了“The study included 92 patients……”；“比较两组……的差异”原来的翻译“The differences in…… were compared”改后直接删除The differences”，更符合英文表达；
4. Results部分，中文原文最后一句“两组比较，差异均有统计学意义”，删除了原文翻译，而是在结果的表述中加入significantly lower/higher，保留句尾（P<0.05），更加符合学术论文中对统计学差异的陈述习惯；
5. Conclusion部分，将can reduce改成了effectively reduces（一般现在时，表示一个确定性的共识结果），时态的改动更符合此处的表达。
当然，这只是摘要的示例，全文翻译也都可以这样，让AI自己翻译自己反思改进。
译文质量大大提升，并且非常符合学术规范，可以说是目前中译英最好的解决方案。
有位发过几十篇文章的老师反馈，单纯中译英的效果不润色直接投也没啥问题：
以上。
课程购买方式
公众号后台回复关键字 6400，查看加入方式
如果你觉得我的分享对你有帮助的话，欢迎大家在这里点赞、在看、分享。当然，也欢迎大家在这里打赏。互动越多，更新越快哦~
声明：本公众号的所有原创内容，在未经允许的情况下，不得用于任何商业用途，违者必究。
