职业，人工智能研究，经验(Career, AI Research, Experience)

Hello everyone! Finally, I can write an article again. There are several ideas that I would want to write. Before, I want to write about another fun project with Deep Learning. Unfortunately, the plan got sidetracked because I got other things to do. After I’m free, I forgot the project and decided to abandon it. Then, I suddenly want to write about my research career experience. Hence, I present you with this article.

^ h ELLO大家！ 最后，我可以再写一篇文章。 我想写几个想法。 在此之前，我想写一个关于深度学习的有趣项目。 不幸的是，该计划被搁置了，因为我还有其他事情要做。 有空之后，我忘记了这个项目，决定放弃它。 然后，我突然想写一下我的研究职业经历。 因此，我向您介绍这篇文章。

In this article, I want to share what I have learned in my job. For your information, I am an AI Research Scientist in Kata.ai. My job desk is to research to tackle some Natural Language Processing (NLP) problems with Machine Learning (ML). It has been one year for me worked as an Artificial Intelligence (AI) Researcher where previously, my job was not a researcher. I would say that I am relatively new in this career.

在本文中，我想分享我在工作中学到的东西。 供您参考，我是Kata.ai中的AI研究科学家。 我的工作台是研究解决机器学习(ML)的某些自然语言处理(NLP)问题。 对我来说，担任人工智能(AI)研究人员已经一年了，以前，我的工作不是研究人员。 我会说我在这个职业中相对较新。

Based on my experience that emphasizes AI’s research environment, I want to share the researcher traits based on my experience working on my research team. The team consists of several Ph.D. graduates. I want to give you an overview of some of the researcher traits. The overview is based on my view, which has just stepped on a researcher career. I hope it provides some insights for anyone who has interested in it.

基于强调AI研究环境的经验，我想根据我在研究团队中的经验来分享研究人员的特征。 该团队由数名博士组成。 毕业生。 我想给您概述一些研究人员的特征。 概述基于我的观点，该观点刚刚涉足研究人员的职业。 我希望它为对它感兴趣的任何人提供一些见识。

批判性思考 (Critical Thinking)

I think this trait is necessary for someone who wants to become a researcher. As you can see, this career’s job is to research something, discover something such as insight based on the research result. The reason is that we need to make sure that the finding is valid. In other words, we want to make people believe that our result is substantial.

我认为此特征对于想要成为研究人员的人来说是必要的。 如您所见，这个职业的工作是研究某些东西，根据研究结果发现一些洞察力。 原因是我们需要确保该发现有效。 换句话说，我们想让人们相信我们的结果是实质性的。

One of the attitudes that I feel about how we do critical thinking is skepticism. It is pretty counter-intuitive, right? Where we usually need optimism on solving a problem, but here we use skepticism. What is it about?

我对我们如何进行批判性思维的态度之一是怀疑主义。 这很违反直觉，对吗？ 我们通常在解决问题时需要乐观的态度，但在这里我们持怀疑态度。 什么事啊

Skepticism in here means that we doubt a research result. For example, when we find some paper that tells that the paper’s finding is promising, We ask this in our mind: “Is it true? I do not think that this approach work on B”.

这里的怀疑态度意味着我们怀疑研究结果。 例如，当我们发现某篇论文表明该论文的发现很有前途时，我们在脑海中问：“这是真的吗？ 我认为这种方法不适用于B”。

This mindset can give a new research idea. You can validate the statement that you doubt, and if it is true, it can be a new research project that want to be solved.

这种思维方式可以给出一个新的研究思路。 您可以验证您怀疑的陈述，如果它是正确的，则可能是一个需要解决的新研究项目。

Before I went to the research team, I was a person who thinks that when that paper has cool intuition and approach, and popular is great. Thanks to my team that I can shed some light on this mindset. Not that I become fully skeptic toward any papers, but at least I got “too good to be true” statement in my brain, especially when I read something that claims great at everything.

在我去研究团队之前，我是一个人，认为那篇论文具有很酷的直觉和方法，并且受欢迎是很棒的。 多亏了我的团队，我可以阐明这种思维方式。 并不是说我对任何论文都持完全怀疑的态度，但至少我脑子里说的“太好了，难以置信”，尤其是当我读到了一篇对所有论文都赞不绝口的文章时。

I have a friend that study for a bachelor’s degree did a research project in NLP. Since he is unfamiliar with it, he decided to ask a Ph.D. student to become his advisor. His advisor gave the paper that has a promising method on it. The advisor suggests my friend use the paper’s approach for his project. He decided to trust the research paper that his advisor gave. Unfortunately, at the end of the project, he failed to create a good model with the paper’s approach, thus cost him a bad grade for that subject. A year later, he realizes that the paper is crap.

我有一个朋友，学士学位，曾在NLP进行过一项研究项目。 由于他不熟悉该技术，因此他决定去攻读博士学位。 学生成为他的顾问。 他的顾问给论文提供了一种很有前途的方法。 顾问建议我的朋友在他的项目中使用论文的方法。 他决定信任其顾问提供的研究论文。 不幸的是，在项目结束时，他未能使用论文的方法来创建一个好的模型，从而使他在该学科上的成绩很差。 一年后，他意识到纸张很烂。

From the story above, if my friend has a critical thinking trait, he can avoid bad grades. Being skeptical about whether the paper is good or not can save you. Nevertheless, skepticism does not mean that we always think negatively onto something, but we assess the object that we pay attention to objectively.

从上面的故事中可以看出，如果我的朋友具有批判性的思维特质，他可以避免成绩不好。 怀疑纸张是否好可以节省您的时间。 尽管如此，怀疑并不意味着我们总是对某事持消极想法，而是我们客观地评估了我们关注的对象。

好奇心 (Curiosity)

With researching as the daily task for researcher, curiosity, in my humble opinion, must present in the mindset of any researcher. With having a curiosity trait, we can enrich the research finding. Thus, making the result more insightful to be digested.

以研究为研究人员的日常工作，我认为，好奇心必须体现在任何研究人员的心态中。 具有好奇心特质，我们可以丰富研究发现。 因此，使结果更具洞察力，易于消化。

When we (my research team) are given a problem, we discuss several ideas on how to solve the problem. The discussion is one of the events where the sense of curiosity of a researcher is needed. We formulate several hypotheses, thinking several approaches might be worth to try to solve the problem.

当我们(我的研究团队)遇到问题时，我们将讨论一些有关如何解决问题的想法。 讨论是需要研究者好奇心的事件之一。 我们提出了几种假设，认为有几种方法可能值得尝试解决该问题。

In my team, when we answer a research question, it usually won’t end with that. This is also where curiosity is needed. Specifically, after we have done implementing and experimenting with a question, we need to analyze the result. Often, several questions pop on their head to explore other things related to the result when analyzing the result. For example, for an AI researcher that focuses on Deep Learning, when he develops a new architecture model that gives significant results, he can be tempted to question its goodness. One of the methods is doing an ablation study, where he can experiment where he can see if the quality degrades when a part of the new model is not present. With curiosity, it can enrich the analysis of the finding, which may further convince the reader.

在我的团队中，当我们回答一个研究问题时，通常不会以此结束。 这也是需要好奇心的地方。 具体来说，在完成问题的实施和实验后，我们需要分析结果。 通常，在分析结果时会突然冒出几个问题，以探索与结果相关的其他事物。 例如，对于专注于深度学习的AI研究人员而言，当他开发出可带来显著成果的新架构模型时，他很容易质疑其优劣。 其中一种方法是进行消融研究，他可以在其中进行实验，以查看当不存在新模型的一部分时质量是否下降。 出于好奇，它可以丰富对发现的分析，从而可以进一步说服读者。

To sum it up, curiosity enhances the research work further. In my opinion, the result of the research works, one of them is shaped by the team’s curiosity. When giving enough curiosity, it will output an insightful finding.

综上所述，好奇心进一步增强了研究工作。 在我看来，研究工作的结果之一是受到团队好奇心的影响。 给予足够的好奇心时，它将输出有见地的发现。

强硬 (Tough)

Not all research experiments output the desired result since a researcher walk to research the unknown. So, be tough when a research result gives terrible results.

由于研究人员开始研究未知数，因此并非所有的研究实验都能输出预期的结果。 因此，当研究结果给出可怕的结果时要坚强。

In my experience, I have experienced many failed experiments. At first, it is unfortunate for me to know that my experiment is fruitless. But, later, I realized that failure is a lesson that can make me a better researcher. The failure experience told me the error which needs to be avoided in the future. What does not kill you make you stronger, right?

以我的经验，我经历了许多失败的实验。 起初，我很不幸地知道我的实验是徒劳的。 但是，后来，我意识到失败是一个可以使我成为更好的研究人员的课程。 失败的经历告诉我以后需要避免的错误。 什么不会杀死你使你变得更强大，对吗？

So yeah, by accepting your failure, it will make you grow. Nevertheless, when a research project fails, it does not mean that all that work is for naught. You can write why you fail in a research paper. It is also an insight for anyone who wants to research on solving a similar problem. Unfortunately, everyone is so fixated on a ‘successful’ research result than the failure one. In my opinion, I think that a failure story is more valuable than the successful one, which can be a warning for someone not to do it again. Though, The written report needs to emphasize the reason for the failure to be valuable thoroughly.

是的，通过接受失败，它会使您成长。 然而，当一个研究项目失败时，这并不意味着所有工作都是徒劳的。 您可以在研究论文中写下失败的原因。 对于任何想研究解决类似问题的人来说，这也是一种见识。 不幸的是，每个人都对一项“成功的”研究结果而不是失败的研究结果如此关注。 我认为，失败的故事比成功的故事更有价值，这可以警告某个人不要再次这样做。 但是，书面报告需要强调未能充分有价值的原因。

So, be tough that a research experiment can give a bad result. Learn and move forward from it!

因此，很难使研究实验产生不好的结果。 从中学习并向前迈进！

系统作家 (Systematic Writer)

After experimenting, we will have a finding or something that we proudly research. So what’s next? Yes, tell the result to everyone about its coolness!

经过实验后，我们将获得一个发现或我们自豪地研究的内容。 下一个是什么？ 是的，将结果告诉所有人有关它的凉爽！

A written report is one of the channels to tell everyone. In my team, we often write a research paper as the form of the report. It is one of our media, where we communicate our finding to the outside.

书面报告是告诉所有人的渠道之一。 在我的团队中，我们经常以报告的形式撰写一份研究论文。 它是我们的媒介之一，我们将我们的发现传达给外界。

When we talk to someone, we know that we need to make sure that our message can be understood. Then, when we want to convince someone that our opinion is reliable, we need to give some reasonable arguments. It is the same as how we write a report. We need to systematically write a reasonable and easy to understand report to convince that our finding is fruitful. Then, the more insightful and easy to follow our writing is, the more people can acknowledge what we research.

当我们与某人交谈时，我们知道我们需要确保我们的信息能够被理解。 然后，当我们想说服某人我们的观点是可靠的时，我们需要给出一些合理的论据。 这与我们编写报告的方式相同。 我们需要系统地编写一份合理且易于理解的报告，以说服我们的发现富有成果。 然后，我们的写作越有洞察力和更容易理解，就会有更多的人认可我们的研究。

I learned here that when we write the paper, we need to write a story to connect our finding systematically. We need to connect the motivation, background, experiment result, and the analysis to make the report. We think that It will make our readers satisfied and enjoy our work.

我在这里了解到，当我们写论文时，我们需要写一个故事来系统地联系我们的发现。 我们需要将动机，背景，实验结果和分析联系起来才能做出报告。 我们认为它将使我们的读者满意并享受我们的工作。

Well.. Actually, the main reason is to pass the reviewer’s judgment on submitting the paper ????.

嗯。实际上，主要原因是通过审稿人对论文????的判断。

结论 (Conclusion)

In this article, I gave several researcher traits that I feel based on my experience working with an AI research team. They are Critical Thinking, Curiosity, Tough, and Systematic Writer. Each of the traits can benefit how they do research projects better.

在本文中，基于与AI研究团队合作的经验，我给出了一些研究人员特质。 他们是批判性思维，好奇心，坚韧和系统的作家。 每个特征都可以使他们更好地进行研究项目。

To be honest, I am still far from being a full-fledged researcher after seeing how a Ph.D. graduates AI researcher work. But I like working with them, especially in discussing how to tackle a problem. I learned a lot from them, especially on how they approach the problem.

老实说，在看到博士学位后，我还远不是一个成熟的研究人员。 毕业于AI研究人员的工作。 但我喜欢与他们合作，尤其是在讨论如何解决问题时。 我从他们那里学到了很多东西，尤其是关于他们如何解决问题的知识。

后记 (Afterwords)

Thank you for reading my article ????. Next time, I want to write something technical on AI. Please give me some strengths to write my next article.

感谢您阅读我的文章article。 下次，我想写一些有关AI的技术文章。 请给我一些力量来写我的下一篇文章。

As always,

一如既往，

I welcome any feedback that can improve myself and this article. I’m in the process of learning on writing and learning to become better. I appreciate any feedback to make me better.

我欢迎任何可以改善自己和本文的反馈。 我正在学习写作和学习变得更好。 感谢您提出的任何改善意见。

See ya in my next article!

在我的下一篇文章中见！