Outline

1. Example of machine Learning
2. Components of Learning
3. Perceptron
4. Types of learning
5. Puzzle

Machine Learning

• Essence
  1. A pattern exists
  2. We cannot pin it down mathematically
  3. we have data on it.

Perceptron

• $h(\mathbf x) = \operatorname{sign} \left( \sum_{i=0}^d w_i x_i \right) = \operatorname{sign} (\mathbf w^T \mathbf x)$

PLA

• steps:
  1. Given the training set: $(\mathbf x_1, \mathbf y_1),(\mathbf x_2, \mathbf y_2), \cdots, (\mathbf x_N, \mathbf y_N)$
  2. pick a misclassified point: $sign (\mathbf w^T \mathbf x_n) \neq y$
  3. update the weight vector: $\mathbf w \leftarrow \mathbf w + y_n \mathbf x_n$

Types of learning

• Supervised learning: input “correct output”.
• Unsupervised learning: no “correct output” input.
• Reinforcement learning: introduce the grade of output.

这门课在证明一件事：通过fit数据，就可以”learn“到未知目标函数。