c语言儿童教学_五岁儿童的自然语言处理

c语言儿童教学

什么是自然语言处理？ (What is Natural Language Processing?)

Don’t be alarmed by the bombastic words! All “natural language processing” means is that we’re trying to help computers to understand human language.

不要被那些夸张的话惊呆了！ 所有“自然语言处理”意味着我们正在尝试帮助计算机理解人类语言。

If you were to go to live in a foreign country, you would have to learn the language, right? Well, this is similar. Computers live with us, and to them, the world is a foreign country and we’re trying to teach them to speak our language. You could teach them to speak any human language, but we’ll talk about teaching them English today.

如果要去国外生活，就必须学习语言，对吗？ 好吧，这是相似的。 计算机与我们同住，对他们来说，世界是一个外国，我们正在尝试教他们说我们的语言。 您可以教他们说任何人类语言，但是今天我们将谈论教他们英语。

Photo by Esther from Pexels

Pexels的 Esther 摄

Computers are really good at math. You could say that math is their language. So to teach computers English, we need to translate English into math. Just like you or I would need an English-Japanese dictionary if we wanted to learn Japanese!

计算机真的很擅长数学。 你可以说数学是他们的语言 。 因此，要教计算机英语，我们需要将英语翻译成数学。 就像您一样，如果我们想学习日语，我将需要一本英语-日语词典！

Smart computer scientists have figured out a way to translate English into computers’ native language (math) by using what they call ‘language models.’

聪明的计算机科学家已经找到了一种方法，可以使用所谓的“语言模型”将英语翻译成计算机的本地语言(数学)。

With these language models, computers are able to do tasks in English just like human beings! For example, your mummy can predict whether or not you will like a future meal depending on your past comments about food! If you often say something like ‘I don’t really like meatloaf,’ your mummy will know that including it in the meal means that you might not enjoy the meal.

使用这些语言模型，计算机可以像人类一样用英语完成任务！ 例如，您的木乃伊可以根据您过去对食物的评论来预测您是否想再吃一顿饭！ 如果您经常说“我不太喜欢肉饼”之类的话，您的木乃伊会知道，将其包含在餐中意味着您可能会不喜欢此餐。

Because computers can now speak English, they can also classify and remember foods that you like and don’t like depending on what you say.

由于计算机现在可以说英语，因此它们还可以根据您说的话来分类和记住您喜欢和不喜欢的食物。

cottonbro from cottonbro，来自 Pexels Pexels

In fact, you’re probably familiar with apps like Postmates or Foodpanda where your parents order food when they do not want to cook. These apps often remember the types of food that your family likes to eat depending on what you guys have expressed about the food in the past!

实际上，您可能熟悉诸如Postmates或Foodpanda之类的应用，您的父母在不想做饭的时候点菜。 这些应用程序经常会记住您家人喜欢吃的食物的类型，具体取决于你们过去对食物的评价！

Some of these kinds of apps can use computers with these English-math dictionaries (language models) to read your parents’ English Language reviews of past restaurants to figure out if you’ll like a new restaurant! Just like your mummy can predict whether or not you’ll like a meal depending on your past comments about food! Super useful, right?

这些类型的应用程序中的某些可以使用带有这些英语数学词典(语言模型)的计算机来阅读父母对过去餐馆的英语评论，从而确定您是否会喜欢新餐馆！ 就像您的木乃伊可以根据您过去对食物的评论来预测您是否喜欢一顿饭一样！ 超级有用吧？

语言模型 (Language Models)

You’re probably curious about how language models work! Well, let's talk about how a language model would ‘translate’ a page in a book from English to math.

您可能对语言模型的工作方式感到好奇！ 好吧，让我们谈谈语言模型如何将书中的页面从英语“翻译”成数学。

Remember that math uses numbers. So a very simple language model would simply label each word that has ever existed with a number. For example, ‘I’ might be word number 0, and ‘am’ might be word number 2. The model could do this for every word in the dictionary.

请记住，数学使用数字。 因此，一个非常简单的语言模型将简单地用数字标记曾经存在的每个单词。 例如，“ I”可能是第0个单词，“ am”可能是第2个单词。模型可以为字典中的每个单词执行此操作。

Photo by Author.

图片由作者提供。

Although this is a successful translation from English to math, computers often find it easier when we help them to understand the relationships between the words in the dictionary, instead of the simple labels we talked about! So we often include information about how often two words occur near each other. To illustrate, you might imagine that ‘Good’ and ‘morning’ occur near each other quite often because we say the phrase ‘Good morning’ a lot! So some language models include this kind of information instead of plain labels.

尽管这是从英语到数学的成功翻译，但是当我们帮助他们理解字典中单词之间的关系时，计算机通常会更轻松，而不是我们谈论的简单标签！ 因此，我们通常会提供有关两个单词彼此相邻出现的频率的信息。 为了说明这一点，您可能会想像“早上好”和“早上好”经常出现在彼此附近，因为我们经常说“早上好”！ 因此，某些语言模型包括此类信息而不是普通标签。

It turns out that when computers know which words occur together often, they can become really smart at problems like the one we discussed earlier — of deciding whether or not your comment about a type of food means you like the food!

事实证明，当计算机知道哪些单词经常一起出现时，它们可以真正解决诸如我们之前讨论的问题-决定您对某种食物的评论是否意味着您喜欢这种食物！

结论 (Conclusion)

I really hope that you liked learning about teaching computers to speak human languages! If you want to learn more, do let me know! You’ve been an excellent student!

我真的希望您喜欢学习教计算机说人类语言！ 如果您想了解更多，请告诉我！ 你是一个优秀的学生！

Gain Access to Expert View — Subscribe to DDI Intel

获得访问专家视图的权限- 订阅DDI Intel

翻译自: https://medium.com/datadriveninvestor/natural-language-processing-for-five-year-olds-b07b832b2de2

c语言儿童教学