# 第一周：深度学习引言(Introduction to Deep Learning)

## 1.1 Supervised Learning with Neural Networks

![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/14.bmp)

一般的**监督式学习**（房价预测和线上广告问题），只要使用标准的神经网络模型就可以

图像识别处理问题，则要使用**卷积神经网络（Convolution Neural Network）**，即**CNN**

处理类似语音这样的序列信号时，则要**使用循环神经网络（Recurrent Neural Network）**，即**RNN**

自动驾驶这样的复杂问题则需要更加复杂的**混合神经网络模型**

![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/15.bmp)

**CNN**一般处理图像问题，**RNN**一般处理语音信号

数据类型一般分为两种：**Structured Data**和**Unstructured Data**

![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/16.bmp)

**Structured Data**通常指的是有实际意义的数据，例如房价预测中的size，#bedrooms，price等；例如在线广告中的User Age，Ad ID等

**Unstructured Data**通常指的是比较抽象的数据，例如Audio，Image或者Text

![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/336import.png)![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/337import.png)

## 1.2 Why is Deep Learning taking off？

![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/17.bmp)

红色曲线代表了传统机器学习算法的表现，例如是SVM，logistic regression，decision tree等。当数据量比较小的时候，传统学习模型的表现是比较好的。当数据量很大的时候，其性能基本趋于水平

构建一个深度学习的流程是首先产生Idea，然后将Idea转化为Code，最后进行Experiment。接着根据结果修改Idea，继续这种Idea->Code->Experiment的循环，直到最终训练得到表现不错的深度学习网络模型

![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/338import.png)![](https://baozou.gitbooks.io/neural-networks-and-deep-learning/content/assets/339import.png)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://baozoulin.gitbook.io/dl-ai-special-notebook/course1/di-yi-zhou-ff1a-shen-du-xue-xi-yin-8a0028-introduction-to-deep-learning-course1-di-yi-zhou-ff1a-shen.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.