玩转ChatGPT:论文翻译润色

这篇具有很好参考价值的文章主要介绍了玩转ChatGPT:论文翻译润色。希望对大家有所帮助。如果存在错误或未考虑完全的地方,请大家不吝赐教,您也可以点击"举报违法"按钮提交疑问。

一、写在前面

首先还是让小Chat推销下自己:

嘿!你是否在写论文的过程中感到头疼,无从下手?你是否在担心自己的语言表达不够专业、不够流畅,影响了论文的质量?不要担心,ChatGPT的润色服务可以帮助你!我们的润色服务包括语言润色、格式排版、逻辑结构调整、学术用语替换等多个方面,旨在帮助您把文章的优点发挥到极致,并提供专业的学术建议和修改建议,让您的文章达到最优的状态。我们的服务简单易用,只需上传您的论文并选择您需要的服务,我们将为您提供快速和高效的润色服务,让您的论文在语言和结构上更加完美,更具有说服力和可读性。同时,我们的服务全天候在线,无论您身在何处,都可以随时享受我们的帮助。让ChatGPT的润色服务帮您更好地表达您的思想,提高您的写作质量,助您在学术之路上更加顺利!

要注意的是,它只是辅助,而且驾驭它是需要技巧的。本篇旨在分享自己的工作流,结合了网络各类大神的总结和自己的实战经验,供参考。

二、实战

摘要:

目的:鉴定马尔尼菲篮状菌(TM)的分泌蛋白并初步分析其生物学功能。

方法:基于质谱技术对 TM 菌株(ATCC18224)培养上清进行分泌蛋白组鉴定利用生物信息学手段预测所鉴定蛋白的亚细胞定位和信号肽剪切位点以及分析其生物学功能和信号通路。

结果:通过2次独立重复实验共鉴定得到21个分泌蛋白功能注释结果显示分泌蛋白主要参与超氧化物代谢、压力应对、金属离子转运等生物学过程预测结果显示分泌蛋白主要定位于细胞核、细胞外、质膜、线粒体和细胞膜且仅有6个蛋白存在信号肽序列。

结论:TM分泌蛋白可能与其免疫逃逸、真菌毒性密切相关且部分蛋白功能尚不明确需要深入研究。

常用工作流:谷歌翻译或者DeepL翻译 >> 小Chat润色。因为小Chat的训练语料毕竟绝大部分是英文,所以TA的中文理解能力其实不太行。

(1)谷歌翻译

谷歌翻译大家都懂:https://translate.google.com.hk/?hl=zh-CN&sourceid=cnhp

我直接上结果:

Summary

Objective: To identify the secreted protein of T. marneffei (TM) and to analyze its biological function.

Methods: Based on mass spectrometry, the secretome identification of TM strain (ATCC18224) culture supernatant was carried out, and the subcellular localization and signal peptide cleavage sites of the identified proteins were predicted by bioinformatics methods, and their biological functions and signaling pathways were analyzed .

Results: Through two independent repeated experiments, a total of 21 secreted proteins were identified; the results of functional annotation showed that the secreted proteins were mainly involved in biological processes such as superoxide metabolism, stress response, and metal ion transport; the predicted results showed that the secreted proteins were mainly located in nucleus, extracellular, plasma membrane, mitochondria and cell membrane, and only 6 proteins had signal peptide sequences.

Conclusion: TM secreted protein may be closely related to its immune escape and fungal toxicity, and the function of some proteins is not yet clear, which needs further study.

(2)DeepL翻译器

这个东西可能有些童鞋没听说过:https://www.deepl.com/translator

看看TA的翻译效果,缺点就是卡:

Abstract

Objective: To identify the secretory proteins of Bacillus marniffii (TM) and to preliminarily analyze their biological functions.

Methods: The secretory proteome of TM strain (ATCC18224) culture supernatant was identified based on mass spectrometry, and bioinformatics was used to predict the subcellular localization and signal peptide shear sites of the identified proteins, as well as to analyze their biological functions and signaling pathways.

Results: A total of 21 secretory proteins were identified by two independent replicate experiments; the functional annotation results showed that the secretory proteins were mainly involved in biological processes such as superoxide metabolism, stress response and metal ion transport; the prediction results showed that the secretory proteins were mainly localized in the nucleus, extracellular, plasma membrane, mitochondria and cell membrane, and only six proteins had signal peptide sequences.

Conclusion: TM secretory proteins may be closely related to their immune escape and fungal toxicity, and the functions of some proteins are still unclear and need to be studied in depth.

(3)谷歌和DeepL的翻译对比

推荐一个文本对比工具,方便对比查看

贴网址:https://www.qianbo.com.cn/Tool/File-Difference/?wl=1

玩转ChatGPT:论文翻译润色

玩转ChatGPT:论文翻译润色

我的品鉴:①第一个谷歌就错了,Abstract嘛,学过的都懂;②马尔尼菲篮状菌的专业英文,DeepL没搞对;③综合来看,DeepL好一些(英语渣渣也只看得出这些了)。

下面用魔法来评价魔法,看看GPT-4的评语:

玩转ChatGPT:论文翻译润色

所以说,我的直觉还是对的,DeepL稍微好一些(不过TA没有点出Abstract的问题,哈哈哈)。所以接下来,我们基于DeepL进行进一步润色。

(4)上小Chat(3.5版本)

(a)咒语一:现在常规驾驭小chat的话术就是“角色扮演法”,就是先让TA扮演特定角色,(偷偷告诉大家:GPT-4拥有强大的角色扮演能力)不过呢,GPT-3.5其实也够用了。

Prompt: You are now acting as an expert in the field of writing SCI academic papers. Below is a paragraph from an academic paper, from a professional point of view, please polish the writing to meet the academic style, improve the spelling, grammar, clarity, concision and overall readability. Be careful not to modify the full text or add any new content, just modify the original sentence. Furthermore, list all modification and explain the reasons to do so in markdown table. The content as follows: […]

Revised content:

Objective: The objective of this study is to identify the secretory proteins of Bacillus marniffii (TM) and analyze their biological functions.

Methods: The secretory proteome of the TM strain (ATCC18224) culture supernatant was identified using mass spectrometry. Bioinformatics tools were employed to predict the subcellular localization and signal peptide cleavage sites of the identified proteins, as well as to analyze their biological functions and signaling pathways.

Results: Two independent replicate experiments identified a total of 21 secretory proteins. Functional annotation revealed that the secretory proteins were primarily involved in biological processes such as superoxide metabolism, stress response, and metal ion transport. The prediction results indicated that the secretory proteins were mainly localized in the nucleus, extracellular space, plasma membrane, mitochondria, and cell membrane. Only six proteins had signal peptide sequences.

Conclusion: The secretory proteins of TM strain may play a crucial role in immune escape and fungal toxicity. However, the functions of some proteins are yet to be studied in depth.

玩转ChatGPT:论文翻译润色

(b)咒语二:纠错[1],比如说TM就翻译错了。

Prompt: Note that it is not “Bacillus marniffii”, but “Talaromyces marneffei”.

Note that it is not “The objective of this study is to identify”, but “To identify”.

这样TA就按照你的要求修改过来了。

其实很多时候,TA的回答并不是令人满意的,所以需要让TA重新生成回答可以在后面添加你认为的要着重修改的方向。

Prompt: Your answer is not good enough. Please answer again, this time focusing on[…].

(c)咒语四:多版本参考[1]。

Prompt: Please provide multiple versions for reference.

玩转ChatGPT:论文翻译润色

提供了三个版本

可以让TA展示三个版本有何区别:

Prompt: List the difference between these three versions.

玩转ChatGPT:论文翻译润色

(d)咒语五:节省空间,只显示修改部分[1]。

Prompt: [Put your requirements here…] , since your output length is limited, in order to save space. Please use ellipses for the parts you don’t think need to be modified.

本例子改动的其实不多了,所以效果不明显,就不展示了。

(e)咒语六:精雕细琢[1]。对某个修改的句子感觉不满意的,让TA给出理由,最终自己定夺。

Prompt: For the sentence “[Before polished sentence]”, why did you polish it to be “[Polished sentence]”.

例如:For the sentence, “A total of 21 secretory proteins were identified by two independent replicate experiments”, why did you polish it to be “Two independent replicate experiments identified a total of 21 secretory proteins.”.

TA开始狡辩,我说换回被动时态。然后,TA开始彩虹屁了。

玩转ChatGPT:论文翻译润色

(5)其他模块化咒语工具箱

以上就是我平时常用的润色工作流,值得注意的是,小Chart属于一种生成式模型,因此输出的结果是概率性的。可以发现,同样的咒语,可能输出不同的句子

所以,使用各种咒语进行微调就很重要,现在搬运一些咒语供替换[1],大家可以根据需求自行添加替换:

(a)风格控制

More precise:选择更精确的词汇,例如使用“generate”代替“produce”或“analyze”代替“look at”。

More concise:消除不必要的词语和短语,使句子更加清晰、直接。

More objective:删除主观性语言,以中立的方式呈现信息。

More specific:提供更具体的细节,以支持论点或想法。

More coherent:确保句子组织良好,逻辑流畅。

More consistent:确保句子所使用的语言和风格与论文的其余部分一致。

More academic:使用学术写作中常用的术语和短语,例如“furthermore”和“thus”。

More formal grammar:使用正确的语法和句法,例如避免句子碎片或跑题的句子。

More nuanced:通过使用词语或短语来传达更复杂或微妙的含义,使句子更具细节。

(b)程度控制

Subtle edits only: 仅对文本进行微调

Minor edits: 进行一些小的编辑

Rephrase for clarity: 改写以提高表达清晰度

Simplify sentence structure: 简化句子结构

Check grammar and spelling: 检查语法和拼写

Enhance flow and coherence: 提高文本流畅度和连贯性

Improve word choice: 改善用词

Revise for style: 为文本调整风格

Significant edits: 进行重大编辑

Restructure content: 重新构建内容

(c)对线的咒语

Prompt: This sentence is too long and complex. Consider breaking it up into multiple shorter sentences.

提示:这句话太长而复杂,考虑将其分解为多个较短的句子。

Prompt: This section seems repetitive. Please rephrase to avoid redundancy.

提示:本节似乎是重复的。请重塑以避免冗余。

Prompt: This phrase seems out of place. Please rephrase to improve clarity.

提示:这句话似乎不合适。请重新措辞以表达更清晰。

Prompt: I have used a passive voice in this sentence. Consider using an active voice instead.

提示:我在这句话中使用了被动语态。考虑改用主动语态。

Prompt: If I wish to publish a paper at a XXX conference, please polish the above content in the style of a XXX article.

(6)关于GPT-4

那么,GPT-4在这个润色上与GPT-3.5有多大区别呢?

仅用咒语一测试,一锤子买卖:

GPT-4 Revised paragraph:

Objective: To identify the secretory proteins of Bacillus marniffii (TM) and to conduct a preliminary analysis of their biological functions.

Methods: The secretory proteome of the TM strain (ATCC18224) culture supernatant was identified through mass spectrometry. Bioinformatics tools were employed to predict the subcellular localization and signal peptide cleavage sites of the identified proteins, as well as to analyze their biological functions and associated signaling pathways.

Results: A total of 21 secretory proteins were identified in two independent replicate experiments. Functional annotation results revealed that these secretory proteins were primarily involved in biological processes such as superoxide metabolism, stress response, and metal ion transport. Prediction results indicated that the secretory proteins were mainly localized in the nucleus, extracellular space, plasma membrane, mitochondria, and cell membrane, with only six proteins possessing signal peptide sequences.

Conclusion: TM secretory proteins may be closely associated with immune evasion and fungal toxicity. The functions of some proteins remain unclear and warrant further in-depth investigation.

然后和GPT-3.5对比:

玩转ChatGPT:论文翻译润色

玩转ChatGPT:论文翻译润色

 除了TM的专业名词没对,其他我觉得挺好的。

三、写在最后

  1. 如果是中译英,推荐DeepL >> GPT3.5(GPT4);如果是英文润色,直接GPT-3.5(GPT-4);DeepL免费版本有使用次数上限,不行只能用Google。
  2. 使用GPT润色,首先使用“角色扮演”咒语,得出一个版本;然后在根据自己的判断,选择不同的“咒语”,例如纠正某一个专业名词、偏向某种风格等。
  3. 记住原则,是TA辅助你,而不是你受限于TA。因此,现阶段奢望一步到位的润色,是不现实的,该看文献还是得看文献,专业背景知识全面了,才能驾驭TA。
  4. 提供一个俄罗斯套娃的思路,就去问小chat:如果我要解决这个问题,我该如何问你?其实就是Auto-GPT的思路,小Chat你自己想。“咒语”如下图[2],自行探索:

玩转ChatGPT:论文翻译润色

帮你们敲下来了:我想请你润色SCI论文,请问我应该如何向你提问才能得到最满意的答案,请提供全面、详细的建议,针对每一个建议请你提供具体的提问范例,注意这些范例都是关于如何向你提问获取做这件事情的建议的,最后根据你所有的建议,在综合提供一个总的提问范例,注意这个范例必须同时体现你所有的建议。

在帮你们问GPT-4:

To get the most satisfactory answers when asking me to help polish your SCI paper, you should follow these guidelines:

(a) Specify the context: Clearly state the topic, field, or subject of your paper, so I can provide appropriate suggestions.

Example question: "My paper is about the application of machine learning algorithms in medical image analysis. Can you give me advice on how to improve my paper?"

(b) Break down the task: Divide your request into smaller, focused questions. This helps me provide targeted assistance on specific areas.

Example question: "Can you help me improve the abstract of my paper?"

(c) Be clear and concise: Make sure your questions are straightforward, and avoid using vague or ambiguous language.

Example question: "Can you help me rephrase this sentence to make it clearer?"

(d) Provide examples: Include relevant excerpts or samples from your paper, so I can better understand the context and offer specific suggestions.

Example question: "Here's a paragraph from my introduction. Can you help me improve its clarity and flow?"

(e) Request feedback on specific aspects: If you have concerns about particular elements of your paper, such as the structure or presentation of data, ask for feedback on those areas.

Example question: "Can you help me improve the organization of my results section?"

(f) Be open to multiple rounds of feedback: Keep in mind that polishing a paper often requires multiple iterations. Feel free to ask follow-up questions based on my suggestions.

Example question: "I've revised my abstract based on your feedback. Can you review it again?"

Following these guidelines, a comprehensive example question could be:

"My paper is about the application of machine learning algorithms in medical image analysis. I need help improving the abstract, introduction, and organization of the results section. Here's the abstract and a paragraph from the introduction. Can you help me improve their clarity and flow? Additionally, can you provide suggestions on how to better present my data in the results section?"

This example question incorporates all the recommendations, making it easier for me to provide targeted and effective assistance

玩转ChatGPT:论文翻译润色

玩转ChatGPT:论文翻译润色

四、搬运目录文章来源地址https://www.toymoban.com/news/detail-449973.html

  1. CSDN博主「江鸟1998」的原创文章,原文链接:https://blog.csdn.net/He_r_o/article/details/129050606。
  2. 知乎答主「史蒂芬」的原创文章,原文链接:https://www.zhihu.com/question/584402332/answer/2982394770?utm_campaign=&utm_medium=social&utm_oi=49199861202944&utm_psn=1631948511548698624&utm_source=qq

到了这里,关于玩转ChatGPT:论文翻译润色的文章就介绍完了。如果您还想了解更多内容,请在右上角搜索TOY模板网以前的文章或继续浏览下面的相关文章,希望大家以后多多支持TOY模板网!

本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若转载,请注明出处: 如若内容造成侵权/违法违规/事实不符,请点击违法举报进行投诉反馈,一经查实,立即删除!

领支付宝红包 赞助服务器费用

相关文章

  • 论文翻译插件推荐

    火山翻译,字节旗下产品。 火山翻译 - 在线翻译 (volcengine.com)   使用方法: 翻译PDF文件:在浏览器上安装火山翻译插件之后,可以将浏览器打开的pdf文件,在其中选中某行文字进行翻译,方便快捷; 翻译网页:右键翻译网页。 安装火山翻译浏览器插件:火山翻译浏览器插件

    2024年02月13日
    浏览(42)
  • Segment Anything【论文翻译】

    https://segment-anything.com/ https://arxiv.org/pdf/2304.02643.pdf 本文由博主 墨理学AI 借助翻译工具对该论文进行了翻译 ​​ 我们介绍Segment Anything(SA)项目:这是一个全新的任务、模型和图像分割数据集。通过在数据收集循环中使用我们高效的模型,我们建立了迄今为止最大的分割数据

    2024年02月14日
    浏览(51)
  • Segment Anything论文翻译,SAM模型,SAM论文,SAM论文翻译;一个用于图像分割的新任务、模型和数据集;SA-1B数据集

    论文链接: https://arxiv.org/pdf/2304.02643.pdf https://ai.facebook.com/research/publications/segment-anything/ 代码连接:https://github.com/facebookresearch/segment-anything 论文翻译: http://t.csdn.cn/nnqs8 https://blog.csdn.net/leiduifan6944/article/details/130080159 本文提出Segment Anything (SA)项目:一个用于图像分割的新任务

    2023年04月19日
    浏览(45)
  • SOF-SLAM论文翻译

    摘要 -同时定位与绘图(SLAM)在计算机视觉和机器人领域中占有重要地位。为了便于分析,传统的SLAM框架采用了强静态世界假设。如何应对动态环境是一个非常重要的问题,越来越受到人们的关注。现有的动态场景SLAM系统要么单独利用语义信息,要么单独利用几何信息,要么以

    2024年02月16日
    浏览(50)
  • OpenAI 生成视频模型 Sora 论文翻译

    视频生成模型作为世界模拟器 本技术报告的重点是 (1) 将所有类型的视觉数据转换为统一表示,以便对生成模型进行大规模训练的方法,以及 (2) 对索拉的能力和局限性的定性评估。 该报告不包括模型和实现细节。 许多先前的工作使用各种方法研究了视频数据的生成建模,包

    2024年02月20日
    浏览(52)
  • 【半监督医学图像分割 2023 】UCMT 论文翻译

    论文题目:Co-training with High-Confidence Pseudo Labels for Semi-supervised Medical Image Segmentation 中文题目:基于高置信度伪标签的联合训练半监督医学图像分割 论文链接:https://arxiv.org/abs/2301.04465 论文代码:https://github.com/Senyh/UCMT 发表时间:2023年1月 论文团队:东北大学福建师范大学阿

    2024年02月06日
    浏览(48)
  • 论文的重复翻译降重 ai写作

    大家好,今天来聊聊论文的重复翻译降重 ai写作,希望能给大家提供一点参考。 以下是针对论文重复率高的情况,提供一些修改建议和技巧,可以借助此类工具: 论文的重复翻译降重方法 在撰写论文时,重复翻译是一个常见的问题。重复翻译不仅增加了论文的篇幅,还可能

    2024年02月02日
    浏览(60)
  • [论文阅读]Visual Attention Network原文翻译

    [论文链接]https://arxiv.org/abs/2202.09741   虽然一开始是被设计用于自然语言处理任务的,但是自注意力机制在多个计算机视觉领域掀起了风暴。然而,图像的二维特性给自注意力用于计算机视觉带来了三个挑战。(1)将图像视作一维序列忽视了它们的二维结构;(2)二次复杂

    2024年02月09日
    浏览(53)
  • 论文降重翻译器英文 神码ai

    大家好,今天来聊聊论文降重翻译器英文 神码ai,希望能给大家提供一点参考。 以下是针对论文重复率高的情况,提供一些修改建议和技巧,可以借助此类工具: 论文降重翻译器:英文论文降重的利器 一、引言 随着全球学术交流的日益频繁,英文论文的数量不断增加。对于

    2024年02月21日
    浏览(60)
  • 英文论文降重翻译器 ai写作

    大家好,今天来聊聊英文论文降重翻译器 ai写作,希望能给大家提供一点参考。 以下是针对论文重复率高的情况,提供一些修改建议和技巧,可以借助此类工具: 英文论文降重翻译器 如果你正在为英文论文的降重问题而苦恼,那么你一定需要一款实用的工具来帮助你。而

    2024年01月25日
    浏览(48)

觉得文章有用就打赏一下文章作者

支付宝扫一扫打赏

博客赞助

微信扫一扫打赏

请作者喝杯咖啡吧~博客赞助

支付宝扫一扫领取红包,优惠每天领

二维码1

领取红包

二维码2

领红包