空间注意力 通道注意力_请让我注意

空间注意力 通道注意力

How do I get your attention? You’re overloaded with information, right? Speed ​​and noise are two characteristics that define our world today and that makes it increasingly difficult for me to get YOUR ATTENTION.

^ h OW做我得到你的注意？ 您信息太多了吧？ 速度和噪音是今天定义我们世界的两个特征，这使我越来越难以引起您的注意。

Social networks, saturation of news, some fake ones, others not that much, emails, messages, … all of this continually consumes us. Don’t you think it is magical that despite that, companies are able to communicate with us and capture a small part of our attention?

社交网络，新闻饱和，一些虚假消息，其他消息，邮件，消息等等，所有这些不断消耗着我们。 尽管如此，您认为公司能够与我们沟通并吸引我们的一小部分注意力，这不是神奇吗？

Let’s talk about AI in Marketing as the enabler to identify potential leads, hyperpersonalize offers and communication, as well as adapt the text and language used in each of the messages contained in company campaigns according to attributes such as age, personal situation or preference of customers.

让我们谈谈营销中的AI作为识别潜在线索，超个性化报价和交流以及根据年龄，个人情况或客户喜好等属性来修改公司活动中包含的每条消息中使用的文本和语言的促成因素。

If you want to learn more, go visit oscargarciaramos.com

如果您想了解更多，请访问 oscargarciaramos.com

Have you heard about clustering or customer segmentation? What about proactive identification of common consumption patterns? Typical, right?

您是否听说过集群或客户细分 ？ 如何主动识别常见的消费模式 ？ 典型的，对吧？

The right ad, with the right message and offer, placed at the right time in the right channel focused on the right audience.

正确的广告 ， 正确的信息和报价 ，在正确的时间投放在正确的渠道上，聚焦于正确的受众 。

However, how much do companies spend on advertising?

但是，公司在广告上花费多少？

他们可以通过AI优化广告支出吗？ (Can they optimize Ad Spend through AI?)

What do we want to know? How much should the advertiser spend on each of their campaigns.

我们想知道什么？ 广告客户应在每个广告系列上花费多少。

In other words, we want to determine the optimal cost, timing and means to carry out advertising campaigns.

换句话说，我们要确定进行广告活动的最佳成本 ， 时间和手段 。

Nowadays, AI systems can help us identify the channel, the message and even adjust the advertising investment plans for each of our advertising campaigns, analyzing their performance and continuously monitoring the different communication channels. And not only that … Thanks to AI we can track not only the performance of an advertiser’s campaigns, but also their competition.

如今，人工智能系统可以帮助我们识别渠道，消息，甚至调整我们每个广告活动的广告投资计划，分析其效果并持续监控不同的沟通渠道。 不仅如此……借助AI，我们不仅可以跟踪广告客户的广告系列的效果，还可以跟踪他们的竞争。

It is what we call “situational awareness”. As markets evolve and constantly change their patterns, AI-based systems can adjust ad spending by milliseconds by changing advertising content and communication channels.

这就是我们所说的“情境意识” 。 随着市场的发展并不断改变其模式，基于AI的系统可以通过更改广告内容和沟通渠道，以毫秒为单位调整广告支出。

There are techniques called MMM (Marketing Mix Modeling) that allow us to quantify the impact of different marketing inputs on Sales or Market share. Thanks to MMM we are able to know the contribution of each of these variables or marketing inputs in sales and what is the estimated investment that we must dedicate to each of them.

有一些称为MMM(营销组合建模)的技术可以使我们量化不同的营销投入对销售或市场份额的影响。 多亏了MMM，我们才能够知道这些变量或营销投入对销售的贡献，以及我们必须为每个变量付出的估计投资。

This technique is based, among others, on ML & AI algorithms such as Multi-Linear Regression. Our dependent variable may be Sales or Market Share, while our independent variables may be distribution, pricing, spending on paper advertising, TV spending, website visitors or promotional spending, among many others.

除其他外，该技术基于ML和AI算法，例如多线性回归。 我们的因变量可能是销售或市场份额，而我们的自变量可能是分布，定价，纸质广告支出，电视支出，网站访问者或促销支出，等等。

Some of these variables do not necessarily have a linear relationship with sales, such as TVspending, that is, they are not directly proportional to the increase in sales (they are non-linear). Why? An advertisement will create awareness in the audience up to a certain point, from which a greater exposure of it would not create a greater awareness because audience will already know enough about the brand.

其中一些变量不一定与销售额具有线性关系，例如电视支出，也就是说，它们与销售额的增长不成正比( 它们是非线性的 )。 为什么？ 广告将在一定程度上提高受众的知名度，因此，广告的更多曝光不会产生更大的知名度，因为受众已经对品牌有了足够的了解。

We will also have to take into account Sales. At MMM, sales must be divided into two components: base sales, or what we will obtain without carrying out any type of advertising; and incremental salts, generated from marketing activities.

我们还必须考虑到Sales 。 在MMM，销售必须分为两个部分： 基本销售 ，或者我们将不进行任何类型的广告而获得的销售 ； 和营销活动产生的增量盐 。

Now, where do we place the ads?

现在，我们在哪里放置广告？

This can be difficult for a human to know. However, thanks to AI we can get to scan the content of thousands and thousands of pages to determine which ads are the most appropriate according to which channel and audience which consumes that page.

这对于人类来说可能很难理解。 但是，借助AI，我们可以扫描成千上万个页面的内容，以确定哪个广告最适合根据哪个页面使用哪个频道和受众。

Let’s dive into the world of …

让我们潜入……的世界

广告制作。 (The creation of commercials.)

What is the perfect content for an advertisement? Can AI guide us in creating content to increase response rates?

广告的最佳内容是什么？ 人工智能可以指导我们创建内容以提高响应率吗？

There are different techniques such as NLG (Natural Language Generation) used by companies such as the Associated Press, based on the creation of text from structured data. This type of techniques would allow us to create analytical reports, blogs and product descriptions with standardized formats.

基于结构化数据的文本创建， 美联社等公司使用了多种技术，例如NLG(自然语言生成) 。 这种类型的技术将使我们能够创建具有标准化格式的分析报告，博客和产品说明。

We can find one clear example with PR 20/20, the marketing agency behind the Marketing Artificial Intelligence Institute, which has reduced the time and production of reports by 80% thanks to NLG.

我们可以在市场人工智能研究所背后的市场代理PR 20/20中找到一个明显的例子，由于NLG的帮助，该报告将报告的时间和制作时间减少了80％。

Companies like Textio are helping companies like Twitter, P&G, Slack or Nestle in creating “augmented writing” by avoiding gender bias or jargon that may be inappropriate when creating content such as job posts.

诸如Textio之类的公司正在帮助Twitter，P＆G，Slack或Nestle之类的公司创建“增强写作”，避免在创建诸如职位等内容时可能不适当的性别偏见或行话。

And not only that … AI has helped us to significantly reduce the number of fraudulent ads. In 2018 they doubled compared to the previous year.

不仅如此……人工智能还帮助我们大大减少了欺诈性广告的数量。 与上一年相比，2018年它们翻了一番。

Get ready for the next November elections in the United States!

为下一次美国11月的选举做准备！

下一个是什么？ (So, what’s next?)

Although we may not appreciate it or be unaware of it, AI is changing the world of advertising. Let’s see it as a great competitive advantage:

尽管我们可能不欣赏它或不知道它，但AI正在改变广告界 。 让我们将其视为巨大的竞争优势：

Let’s lessen the frustration with more relevant and effective ads.

让我们用更相关，更有效的广告来减轻挫败感。

If we do this, we will improve the media spending strategy, the decision-making capacity and it will mean an increase in the sales of our clients.

如果这样做，我们将改善媒体支出策略，决策能力，这将意味着增加客户的销售额。

Feel free to leave a comment or share this post. Follow me for future post.

随时发表评论或分享这篇文章。 跟 我 来以后的帖子。

And if you want to learn even more, you can find me at oscargarciaramos.com

如果您想学习更多，可以在 oscargarciaramos.com上 找到我。

翻译自: https://towardsdatascience.com/may-i-have-your-attention-please-9c03de3c1933

空间注意力 通道注意力