Lecture 3: The Perceptron

Loading [MathJax]/jax/output/HTML-CSS/fonts/TeX/fontdata.js

Lecture 3: The Perceptron

Assumptions

Binary classification (i.e. $y_i \in \{-1, +1\}$ )
Data is linearly separable

Classifier

$h(x_i) = sign(\vec{w}^\top \vec{x_i} + b)$

$b$ is the bias term (without the bias term, the hyperplane that

$\vec{w}$ defines would always have to go through the origin). Dealing with

$b$ can be a pain, so we 'absorb' it into the feature vector

$\vec{w}$ by adding one additional constant dimension. Under this convention,

$\vec{x_i} \hspace{0.1in} \text{becomes} \hspace{0.1in} \begin{bmatrix} \vec{x_i} \\ 1 \end{bmatrix} \\ \vec{w} \hspace{0.1in} \text{becomes} \hspace{0.1in} \begin{bmatrix} \vec{w} \\ b \end{bmatrix} \\$ We can verify that

$\begin{bmatrix} \vec{x_i} \\ 1 \end{bmatrix} \cdot \begin{bmatrix} \vec{w} \\ b \end{bmatrix} = \vec{w}^\top \vec{x_i} + b$ Using this, we can simplify the above formulation of

$h(x_i)$ to

$h(x_i) = sign(\vec{w}^\top \vec{x})$ Observation: Note that

$y_i(\vec{w}^\top \vec{x_i}) > 0 \Longleftrightarrow x_i \hspace{0.1in} \text{is classified correctly}$ where 'classified correctly' means that

$x_i$ is on the correct side of the hyperplane defined by

$\vec{w}$ . Also, note that the left side depends on

$y_i \in \{-1, +1\}$ (it wouldn't work if, for example

$y_i \in \{0, +1\}$ ).

Perceptron Algorithm

Now that we know what the

$\vec{w}$ is supposed to do (defining a hyperplane the separates the data), let's look at how we can get such

$\vec{w}$ .

Perceptron Algorithm

Geometric Intuition

Below we can see that the point

$\vec{x_i}$ is initially misclassified (left) by the hyperplane defined by

$\vec{w_t}$ at time

$t$ . The middle portion shows how the perceptron update changes

$\vec{w_t}$ to

$\vec{w_{t+1}}$ . Right shows that the point is no longer misclassified after the update.

Perceptron Convergence

Suppose that

$\exists \vec{w}^*$ such that

$y_i(\vec{w}^* \cdot \vec{x}) > 0$

$\forall (\vec{x}_i, y_i) \in D$ .

Now, suppose that we rescale each data point and the

$\vec{w}^*$ such that

$||\vec{w}^*|| = 1 \hspace{0.3in} \text{and} \hspace{0.3in} ||\vec{x_i}|| \le 1 \hspace{0.1in} \forall \vec{x_i} \in D$ The Margin of a hyperplane,

$\gamma$ , is defined as

$\gamma = min_{(\vec{x_i}, y_i) \in D}\{\vec{w}^{*T} \cdot \vec{x_i} \}$ We can visualize this as follows

All inputs $\vec{x_i}$ live within the unit sphere
$\gamma$ is the distance from the hyperplane (blue) to the closest data point
$\vec{w}$ lies on the unit sphere

Theorem: If all of the above holds, then the perceptron algorithm makes at most

$1 / \gamma^2$ mistakes.

Keeping what we defined above, consider the effect of an update on

$\vec{w}$ on

$\vec{w}^\top \vec{w}$ and

$\vec{w}^\top \vec{w}^*$

Consider $\vec{w}^\top \vec{w}^*$ : $(\vec{w} + y\vec{x})^\top \vec{w}^* = \vec{w}^\top \vec{w}^* + y(\vec{x}\cdot \vec{w}^*) \ge \vec{w}^\top \vec{w}^* + \gamma$ The inequality follows from the fact that, for $\vec{w}^*$ , the distance from the hyperplane defined by $\vec{w}^*$ to $\vec{x}$ must be at least $\gamma$ .

This means that for each update, $\vec{w}^\top \vec{w}^*$ grows by at least $\gamma$ .
Now consider →w⊤→w: (→w+y→x)⊤(→w+y→x)=→w⊤→w+2y(→w⊤⋅→x)+y2(→x⋅→x)≤→w⊤→w+1 The inequality follows from the fact that
- $2y(\vec{w}^\top\vec{x}) < 0$ as we had to make an update, meaning $\vec{x}$ was misclassified
- $y^2(\vec{x}^\top \vec{x}) \le 1$ as $y^2 = 1$ and all $\vec{x}$ lives in the unit sphere

This means that for each update,

$\vec{w}^\top \vec{w}$

grows by at most 1

Now we can put together the above findings. Suppose we had $M$ updates. $\begin{align} M\gamma &\le \vec{w}^\top\cdot\vec{w}^* &&\text{By (1)} \\ &= |\vec{w}^\top\cdot\vec{w}^*| &&\text{By (1) again (the dot-product must be non-negative because the initialization is 0 and each update increases it by at least $\gamma$)} \\ &\le ||\vec{w}|| ||\vec{w}^*|| &&\text{By Cauchy-Schwartz inequality} \\ &= ||\vec{w}|| &&\text{As $||\vec{w}^*|| = 1$} \\ &= \sqrt{\vec{w}^\top \vec{w}} && \text{by definition of $\|\vec{w}\|$} \\ &\le \sqrt{M} &&\text{By (2)} \\ &\Rightarrow M\gamma \le \sqrt{M} \\ &\Rightarrow M^2\gamma^2 \le M \\ &\Rightarrow M \le \frac{1}{\gamma^2} \end{align}$

History

Initially, huge wave of excitement ("Digital brains")
Then, contributed to the A.I. Winter. Famous counter-example XOR problem (Minsky 1969):

If data is not linearly separable, it loops forver.