你的浏览器不支持本网站所要求的功能, 现在看到的是本网站简化版本.

为了获得最佳体验, 请使用最新的Chrome, Safari, Firefox 或 Edge 浏览器.

2. 机器学习与深度学习

2.1 机器学习的概念与原理


Powered by impress.js
Ver. 2408

### 机器学习 - 教计算机从数据中学习 - 做出决定或预测 - 每个任务都没有明确的编程 - 随着经验的积累自动改进 - 应用于各种任务和领域
![Machine Learning](img/c02/machine-learning.png) #### Machine Learning
![Machine Learning](img/c02/machine-learning.webp) #### Machine Learning
### 机器学习原理 - 输入数据: 例子, 特征和标签 - 学习算法: 处理数据 - 模型: 学习过程的结果 - 对新数据进行预测或决策
### 机器学习为什么重要 - 实现复杂任务的自动化 - 提高效率和准确性 - 适应新的数据和情况 - 分析和发现大型数据集的模式
![Machine Learning](img/c02/machine-learning-applications.png) #### Machine Learning
### 重要人物 - Arthur Samuel: Machine Learning - Geoffrey Hinton: Deep Learning - Yann LeCun: CNN - Yoshua Bengio: RNN and LSTM - 吴恩达: Coursera and Google Brain - Ian Goodfellow: GAN - Jürgen Schmidhuber: Reinforcement Learning - 李飞飞: ImageNet
![Arthur Samuel](img/c02/author-samuel.webp) #### Arthur Samuel
![Andrew Ng](img/c02/andrew-ng.webp) #### Andrew Ng
![Ian Goodfellow](img/c02/Ian-Goodfellow.jpg) #### Ian Goodfellow
![Jürgen Schmidhuber](img/c02/jurgen-schmidhuber.webp) #### Jürgen Schmidhuber
![Fei-fei Li](img/c02/fei-fei-li.jpg) #### Fei-fei Li
### 重点事件 - 1957: 感知机 Perceptron - 1986: 神经网络的反向传播算法 Backpropagation - 1997: 支持向量机 SVMs - 2006: 深度学习 Deep learning - 2012: AlexNet - 2014: 生成对抗网络 GAN - 2017: Transformer - 2018: Diffusion Models - 2022: ChatGPT
![Backpropagation](img/c02/backpropatation.webp) #### Backpropagation
![SVM](img/c02/svm.png) #### SVM
![GAN](img/c02/gan.jpg) #### GAN
![Transformer Model](img/c02/Transformer-model.jpg) #### Transformer Model
![Diffusion Models](img/c02/diffusion-model.png) #### Diffusion Model
![ChatGPT](img/c02/chatgpt.jpg) #### ChatGPT
### 关键论文 ###### Rosenblatt, F. (1957). [The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain.](https://1-1256632535.cos.ap-beijing.myqcloud.com/res/ita/rosenblatt-1957.pdf) ###### Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). [Learning representations by back-propagating errors.](https://1-1256632535.cos.ap-beijing.myqcloud.com/res/ita/rumelhart-1986.pdf) ###### Cortes, C., & Vapnik, V. (1995). [Support-vector networks.](https://1-1256632535.cos.ap-beijing.myqcloud.com/res/ita/cortes-1995.pdf) ###### Hinton, G. E., Osindero, S., & Teh, Y. W. (2006). [A fast learning algorithm for deep belief nets.](https://1-1256632535.cos.ap-beijing.myqcloud.com/res/ita/hinton-2006.pdf) ###### Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). [ImageNet classification with deep convolutional neural networks.](https://1-1256632535.cos.ap-beijing.myqcloud.com/res/ita/krizhevsky-2017.pdf) ###### Goodfellow, I., Pouget-Abadie, ... & Bengio, Y. (2014). [Generative adversarial networks.](https://1-1256632535.cos.ap-beijing.myqcloud.com/res/ita/goodfellow-2014.pdf)
### 机器学习分类 - 监督学习 Supervised Learning - 无监督学习 Unsupervised Learning - 强化学习 Reinforcement Learning
![Supervised Learning](img/c02/supervised-learning.jpg) #### Supervised Learning
![Unsupervised Learning](img/c02/unsupervised-learning.png) #### Unsupervised Learning
![Reinforcement Learning](img/c02/reinforcement-learning.png) #### Reinforcement Learning
### 监督学习 - 从标记的数据中学习 - 分类和回归 - Linear Regression, SVM, Decision Tree
![Linear Regression](img/c02/linear_regression.png) #### Linear Regression
![Decision Tree](img/c02/decision-tree.png) #### Decision Tree
### 无监督学习 - 从未标记的数据中学习 - 聚类和降维 - K-means, PCA
![K-means](img/c02/k-means.png) #### K-means
![PCA](img/c02/pca.png) #### PCA
### 强化学习 - 从与环境的互动中学习 - 智能体, 环境, 行动和奖励 - Q-learning, Deep Q-Networks, Policy Gradients
![ita-2-1 mindmap](img/c02/mindmap-2-1.png)
### 2.1 机器学习的概念与原理 - 什么是机器学习? 机器学习有哪些主要类型? - 为什么要使用机器学习? - 在机器学习的主要类型中, 列出至少一种常用算法. - 你知道哪些机器学习的应用? ---- [ 1.4 人工智能的发展现状与应用 ](ita-1-4.html) [| Exercises |](ita-exec.html) [ 2.2 神经网络的诞生与发展 ](ita-2-2.html)

黑公网安备23010302001726号   黑ICP备2024033110号-1