aiml_您应该在本周（7月18日）阅读有趣的AIML文章

分析与意见 (ANALYSIS AND OPINION)

AI tech and systems are developing at a rapid rate, new inventions and techniques are presented every week. And yet, people still aren’t satisfied with the level of AI we have currently.

人工智能技术和系统发展Swift，每周都有新发明和新技术出现。 但是，人们仍然对我们目前拥有的AI水平不满意。

Find out why AI hasn’t met our expectations and also how identity theft has evolved to leverage deepfakes and facial recognition data.

找出AI为什么没有达到我们的期望，以及身份盗用如何演变成利用Deepfake和面部识别数据的方法。

本周版本的封面文章介绍： (This week’s edition cover article that presents:)

• Tips and advice on how you can write your data science-focused article on Medium

  关于如何在Medium上撰写以数据科学为重点的文章的提示和建议

• How innovative and cheap identity theft has become in China.

  创新和廉价的身份盗窃在中国已经变得如何。

• The lack of appreciation for current levels of AI

  缺乏对当前AI水平的赞赏

• A neural network architecture that you might not have heard about

  您可能没有听说过的神经网络架构

我如何撰写数据科学博客作者Rebecca Vickery (How I Write a Data Science Blog By Rebecca Vickery)

If you are reading this article, then you are no stranger to Medium, you are also very likely, someone who is interested in AI or has a job related to AI. Probably at some point, you have wanted to write articles on Medium.

如果您正在阅读本文，那么您对Medium并不陌生，您很有可能对AI感兴趣或从事与AI相关的工作。 大概在某个时候，您想写有关Medium的文章。

Well, Rebecca Vickery has written an article that breaks down the process by which she uses to write her articles.

好吧， 丽贝卡·维格里 ( Rebecca Vickery )撰写了一篇文章，详细介绍了她用来撰写文章的过程。

Why should you even take the advice and tips the Rebecca shares within her article? Rebecca has been writing on Medium for two years and has consistently published one to two articles a week. And if that’s not enough, she’s also currently considered a top writer on Medium on the topics Education, Technology and Artificial Intelligence.

为什么您甚至应该接受Rebecca在她的文章中分享的建议和技巧？ 丽贝卡(Rebecca)在《 Medium》上写了两年书，每周持续发表一到两篇文章。 如果那还不够的话，她目前还被认为是“教育”，“技术”和“人工智能”主题的“中型”最佳作家。

So how does Rebecca write data science articles?

那么，丽贝卡如何撰写数据科学文章？

According to Rebecca, she tends to associate her writing career with a goal and purpose. Namely, the purpose for Rebecca is to teach others about data science and at the same time, reinforce her knowledge. This is a purpose that anyone can adopt as machine learning is an ever-growing field with new techniques and tools emerging daily.

根据丽贝卡的说法，她倾向于将写作生涯与目标联系起来。 也就是说，丽贝卡的目的是向他人传授有关数据科学的知识，同时加强她的知识。 这是任何人都可以采用的目的，因为机器学习是一个不断发展的领域，每天都有新的技术和工具出现。

Writing data science articles can be tedious and requires a sturdy process, so how does Rebecca stay motivated?

撰写数据科学文章可能很乏味并且需要一个坚固的过程，那么丽贝卡如何保持动力呢？

She points out in her article that she’s not motivated by monetary incentives or social acknowledge. According to Rebecca, by not suing financial or social validation as a source of motivation, you can have a successful and long writing career.

她在文章中指出，她不受金钱激励或社会认可的激励。 根据丽贝卡所说，通过不诉诸财务或社会验证作为动机，您可以拥有成功而漫长的写作生涯。

Monetary and social verifications do not come quickly during the writing journey. Therefore you need something that can motivate you to keep writing for years.

在写作过程中，不能很快进行货币和社会核查。 因此，您需要一些能够激发您继续写作的东西。

The best advice I got from Rebecca’s article is her article refinement technique and process. Rebecca doesn’t sit down and churn out curation worthy article. She takes several iterations at an article. Each iteration takes the writing piece from a list of ideas to explanations around the ideas…to a full article.

我从丽贝卡的文章中得到的最佳建议是她的文章改进技术和过程。 丽贝卡(Rebecca)不会坐下来，发表有价值的文章。 她在一篇文章中进行了多次迭代。 每次迭代都会从构想清单到围绕构想的解释……直至整篇文章。

The key thing here to the iteration technique is that Rebecca writes anywhere and at anytime.

迭代技术的关键在于Rebecca可以随时随地进行编写。

Give this article a read if you want to be inspired to start writing on Medium. Perhaps shortly, I might be reading one of your articles.

如果您想启发自己开始在Medium上写作，请阅读本文。 也许不久之后，我可能正在阅读您的一篇文章。

本文非常适合： (This article is excellent for:)

• Data Science Bloggers

  数据科学博客

• Data Science Practitioners

  数据科学从业者

您只需$ .07美元就可以在中国的黑市上购买随机的人脸识别照片，作者： 戴夫·格什戈恩 ( Dave Gershgorn) 。 (You Can Buy A Random Facial Recognition Photo On China’s Black Market For Just $.07 by Dave Gershgorn.)

I seem to be reading an article or two from Dave Gershgorn every week, and I think everyone within the AI Industry should too. Dave covers relevant and current topics that are AI and Technology related; this is a healthy dose of contemporary information to balance out all the research and technical Meidum articles you probably read weekly.

我似乎每周都要阅读Dave Gershgorn的一两篇文章，我认为AI行业的每个人也应该阅读。 Dave涵盖与AI和技术相关的相关主题和当前主题； 这是健康的当代信息，可以平衡您可能每周阅读的所有Meidum研究和技术文章。

What could identity theft look like in the next ten years?

未来十年，身份盗用会是什么样？

Well, Dave’s article covers a recent development in one of the many misuses of advanced AI-based technology in China. Dave writes on the emerging sales in China’s black market of digital packages that contain images of faces and accompanying personal data.

好吧，戴夫(Dave)的文章涵盖了中国基于先进AI技术的许多误用之一的最新发展。 戴夫(Dave)在中国黑市的数字化包装销售中写道，其中包含面部图像和附带的个人数据。

These sold packages are all the requirements to bypass a lot of facial authentication systems in many widely used financial, business and lifestyle applications in China.

这些已售出的套餐是在中国许多广泛使用的金融，商业和生活方式应用中绕过许多面部认证系统的全部要求。

From the title of Dave’s article, you can see how ridiculously cheap obtaining the data is.

从Dave的文章标题中，您可以看到获取数据多么便宜。

Dave writes about an advanced package that makes this whole story more bizarre. With an up-payment by a few dollars, you can get a ‘deepfake like’ software that can mobilize the content of still images to mimic subtle head movements.

戴夫(Dave)讲述了一个高级软件包，它使整个故事变得更加离奇。 只需支付几美元的预付款，您就可以得到一个类似'deepfake'的软件，该软件可以动员静止图像的内容来模拟微妙的头部运动。

A couple of years ago all that Dave covers in his article would be in the script of a movie set in the year 2079 covering an advance criminal group. But we see this unfold in real-time in 2020.

几年前，戴夫(Dave)在他的文章中所涵盖的全部内容都是一部2079年的电影剧本，其中涵盖了一个犯罪集团。 但是我们看到这种情况会在2020年实时发生。

Dave also mentions how this method of identity theft might not be so easily applicable to the systems within the US. He also includes that IPhones have an advanced process of facial recognition that measures the depth of facial features.

Dave还提到了这种身份盗用方法可能不太容易适用于美国境内的系统。 他还指出，iPhone具有先进的面部识别流程，可测量面部特征的深度。

But is it only a matter of time before someone becomes a digital version of you for less than $1.

但是，有人以不到1美元的价格成为您的数字版本只是时间问题。

这篇文章非常适合阅读： (This article is a great read for:)

• Futurist and Technologist: Get a brief overview of how illegal activities and crimes are evolving with advance technology and AI.

  未来主义者和技术专家 ：简要了解随着先进技术和AI在非法活动和犯罪方面的演变。

• Machine Learning Practitioners: Get clued on how the misuse of tools and application developed by some of the smartest minds in the world can contribute to a serious problem.

  机器学习从业者 ：了解世界上一些最聪明的人开发的工具和应用程序的滥用如何导致严重问题。

AI革命就在这里。 这与我们对Tobia Tudino的期望不同 (The AI Revolution Is Here. It’s Just Different Than We Expected By Tobia Tudino)

We wanted flying cars and they gave us DeepFakes.

我们想要飞行汽车，他们给了我们DeepFakes。

Tobia Tudino latest article is based on the non-realization of how revolutionary that the current state of AI is. This lack of appreciation is a result of high expectations due to the presentation of AI in science fiction.

Tobia Tudino的最新文章基于未意识到AI当前状态的革命性。 缺乏欣赏是由于在科幻小说中出现了AI导致人们寄予厚望。

Tobia states that the futuristic depiction of the several forms that AI can take in which we observe in sci-fi films and moves has occluded our view from recognizing that the AI revolution we’ve long awaited is here, in front of us.

Tobia指出，在科幻电影和动作中观察到的AI可以采用的几种形式的未来主义描述，使我们无法认识到我们期待已久的AI革命就在眼前。

But I see no flying cars; no teleportation…we do have some pretty cool spacecraft.

但是我看不到飞行的汽车。 没有隐形眼镜……我们确实有一些很酷的 航天器 。

Tobia starts to explain the reason why AI doesn’t meet our cinematic driven expectation with an expanded statement made by Micheal I. Jordan(Artificial Intelligence — The Revolution Hasn’t Happened Yet”). The statement relays the message that AI is a wildcard term. The term “AI” is used to describe any advancement in technology that eliminates or emulates human intervention or involvement.

Tobia用Micheal I. Jordan(《 人工智能—革命还没有发生》 )的扩展声明开始解释AI无法满足我们对电影的期望。 该语句中继消息AI是通配符术语。 术语“ AI”用于描述消除或模仿人类干预或参与的任何技术进步。

And based on that description of AI, it is safe to say we have AI all around us, everywhere.

基于对AI的描述，可以肯定地说我们到处都有AI。

I get the impression that Tobia wants us, the reader, to appreciate the level of AI we have now. Tobia mentions advances in decision making and data gathering within the healthcare, business, consultancy, fashion industries.

我给人的印象是Tobia希望我们的读者欣赏我们现在拥有的AI水平。 Tobia提到了医疗保健，商业，咨询，时尚行业在决策和数据收集方面的进步。

Maybe people are fearful that the appreciation of current AI might lead to content. Or perhaps it might not be such a bad thing that people have yet to accept the fact that we’ve achieved “AI” that we can proud of, as this can push researchers, engineers and innovators to invent and create better AI systems and tech.

也许人们担心当前AI的升值可能会带来满足感。 也许这并不是一件坏事，以至于人们尚未接受我们引以为傲的“ AI”这一事实，因为这会促使研究人员，工程师和创新者发明和创造更好的AI系统和技术。

本文是令人兴奋的阅读文章： (This article is an exciting read for:)

• Machine Learning Practitioners: This article will encourage you yo appreciate the work that you are putting out into the world, even if the world might not necessarily show appreciation.

  机器学习从业者 ：即使您未必表示赞赏，本文也会鼓励您欣赏您正在向全世界推出的工作。

有些人称其为天才，另一些人称其为愚蠢： 安德烈·叶(Andre Ye)创造的最具争议的神经网络 (Some Call it Genius, Others Call it Stupid: The Most Controversial Neural Network Ever Created By Andre Ye)

I’ve never come across the topic of ‘Extreme Learning Machine(ELM)’, so Andre Ye’s article for me was both informative and educational.

我从来没有遇到过“极限学习机(ELM)”这一主题，因此， 安德烈·叶 ( Andre Ye )为我撰写的文章既内容丰富又具有教育意义。

Some of you might also not have heard of ELM, and Andre provides a reason for this in his article. The main reason is along the lines that the machine learning community not adopting it and deep learning experts criticizing and questioning its performance.

你们中有些人可能还没有听说过ELM，Andre在他的文章中提供了这样做的原因。 主要原因是机器学习社区未采用它以及深度学习专家批评和质疑它的性能。

Andre’s article is very educative as it introduces ELM at an elementary level, so anyone with an essential background in Machine Learning can understand what the description and benefits of ELM.

Andre的文章非常有教育意义，因为它在基础级别上介绍了ELM，因此具有机器学习基础知识的任何人都可以了解ELM的描述和优点。

Andre provides an introduction to ELM by stating its internal components; how the neural network architecture is trained and other more specific characteristics and properties.

Andre通过陈述其内部组件来介绍ELM。 如何训练神经网络架构以及其他更具体的特性和属性。

The general idea I got from the article is that ELMs are effective as a result of its random nature. As stated by Andre, ELMs are composed of two lawyers where the first layer is randomly initialized, and the weights parameters are fixed.

我从本文中得出的一般想法是，ELM由于其随机性而有效。 如Andre所述，ELM由两名律师组成，其中第一层是随机初始化的，权重参数是固定的。

Incorporating elements of randomness in neural network architecture is something we see in a variety of deep learning neural network architecture; weights can be randomly initialized, and dropout later also provide randomness characteristics to an architecture.

我们在各种深度学习神经网络架构中看到了将随机性元素纳入神经网络架构的情况。 权重可以被随机初始化，并且以后的丢弃也可以为架构提供随机性特征。

Andre also covers the main reasons why ELMs are dismissed by reputable individuals within the machine learning communities. Andre also includes some information on the limitation faced by ELMs and reasons as to why they are not widely adopted or utilized.

Andre还介绍了ELM被机器学习社区中的知名人士解雇的主要原因。 Andre还提供了一些有关ELM所面临的局限性的信息，以及为何未广泛采用或利用它们的原因。

这篇文章非常适合阅读： (This article is a great read for:)

• Machine Learning Practitioners

  机器学习从业者

希望您觉得这篇文章有用。 (I hope you found the article useful.)

To connect with me or find more content similar to this article, do the following:

要与我联系或查找更多类似于本文的内容，请执行以下操作：

1. Subscribe to my YouTube channel for video contents coming soon here

   订阅我的YouTube频道以获取即将在这里 播出的视频内容

2. Follow me on Medium

   跟我来中

3. Connect and reach me on LinkedIn

   在LinkedIn上联系并联系我