Unsupervised Learning: Difference between revisions

From David's Wiki
Line 82: Line 82:
We will fix <math>\theta</math> and optimize wrt <math>Q</math>.<br>
We will fix <math>\theta</math> and optimize wrt <math>Q</math>.<br>
Jensen's inequality holds with equality iff either the function is linear or if the random variable is degenerate.<br>
Jensen's inequality holds with equality iff either the function is linear or if the random variable is degenerate.<br>
Since log is not linear, we will assume <math>\frac{P(x^{(i)}, z^{(i)}=j; \theta)
We will assume <math>\frac{P(x^{(i)}, z^{(i)}=j; \theta)
}{Q^{(i)}_{(j)}}</math> is a constant.<br>
}{Q^{(i)}_{(j)}}</math> is a constant.<br>
This implies <math>Q^{(i)}(j) = c * P(x^{(i)}, z^{(i)} = j ; \theta)</math>.<br>
This implies <math>Q^{(i)}(j) = c * P(x^{(i)}, z^{(i)} = j ; \theta)</math>.<br>

Revision as of 18:50, 14 November 2019

Basics of Unsupervised Learning

Clustering

Given points \(\displaystyle \{x^{(1)},...,x^{(m)}\}\) we wish to group them into clusters 1,...,k.
Usually we are given the number of clusters, k.
The problem is to assign labels \(\displaystyle \{z^{(1)},...,z^{(m)}\}\) where \(\displaystyle z^{(i)} \in \{1,...,k\}\)

K-means

Here we wish to cluster them to minimize the distance to the mean of their cluster.
In our formulation we have k clusters.
The mean of each cluster \(\displaystyle \mu_i\) is called the centroid.</math>

Optimization

Let \(\displaystyle \mathbf{\mu}\) denote the centroids and let \(\displaystyle \mathbf{z}\) denote the cluster labels for our data.
Our loss function is \(\displaystyle L(\mu, \mathbf{z}) = \sum_{i} \Vert x^{(i)} - \mu_{z^{(i)}} \Vert ^2\).

Using coordinate-block descent

\(\displaystyle L(\mu, \mathbf{z}) = \sum_{i} \Vert x^{(i)} - \mu_{z^{(i)}} \Vert ^2\).
If we fix \(\displaystyle \mu\) then we have only need to minimize \(\displaystyle z^{(i)}\).
Since each term of the sum is independent, we simply choose the closest centroid for each point
If we fix \(\displaystyle \mathbf{z}\) then we need to minimize \(\displaystyle L(\mu, \mathbf{z})\) wrt \(\displaystyle \mu\).
Taking the gradient and setting it to 0 we get:
\(\displaystyle \nabla_{\mu} L(\mu, \mathbf{z}) = \nabla_{\mu} \sum_{i} \Vert x^{(i)} - \mu_{z^{(i)}} \Vert ^2 \)
\(\displaystyle = \nabla_{\mu} \sum_{j=1}^{k} \sum_{i\mid z(i)=j} \Vert x^{(i)} - \mu_{z^{(i)}} \Vert ^2 \)
\(\displaystyle = \nabla_{\mu} \sum_{j=1}^{k} \sum_{i\mid z(i)=j} \Vert x^{(i)} - \mu_{j} \Vert ^2 \)
\(\displaystyle = \sum_{j=1}^{k} \sum_{i\mid z(i)=j} \nabla_{\mu} \Vert x^{(i)} - \mu_{j} \Vert ^2 \)
\(\displaystyle = -\sum_{j=1}^{k} \sum_{i\mid z(i)=j} 2(x^{(i)} - \mu_{j}) = 0 \)
\(\displaystyle \implies \mu_{j} = (\sum_{i\mid z(i)=j} x^{(i)})/(\sum_{i\mid z(i)=j} 1) \quad \forall j \)

Notes

This procedure will yield us a sequence of parameters and losses.
\(\displaystyle \mu^{0}, z^{0}, \mu^1, z^1, ...\)
\(\displaystyle L(0) \geq L(1) \geq L(2) \geq ...\)

  • Since the loss is monotone decreasing and bounded below, it will converge by the monotone convergence theorem.
  • However, this does not imply that the parameters \(\displaystyle \mathbf{\mu}\) and \(\displaystyle \mathbf{z}\) will converge.

Algorithm

  1. Randomly initialize labels \(\displaystyle \mathbf{z}\).
  2. Then calculate the centroids \(\displaystyle \mathbf{\mu}\).
  3. Then update the labels for each example to the closest centroid.
  4. Update the centroids by taking the mean of every point in the cluster.
  5. Repeat steps 3 and 4

Soft K-means

We will develop a model for how our data is generated.
We will then find probabilities for each element being from a cluster (ala Bayesian paradigm).
Given \(\displaystyle k\) clusters, the probability of a point being from cluster k is \(\displaystyle \phi_k = P(z^{(i)} = k)\)
We will assume each cluster is from a normal distribution \(\displaystyle N(\mu_j, \sigma_j)\)

EM Algorithm

Expectation Maximization
The key idea is to introduce intermediate coefficients to apply Jensen's inequality.

Maximum Likelihood

Let \(\displaystyle \theta_j = [\phi_j, \mu_j, \sigma_j]\) denote our parameters.
The likelihood is \(\displaystyle L(\theta) = \Pi_{i=1}^{m} P(x^{(i)}; \theta)\) where \(\displaystyle P\) is our pdf/pmf.
Then the log-likelihood is \(\displaystyle l(\theta) = \sum_{i=1}^{m} \log P(x^{(i)}; \theta) = \sum_{i=1}^{m} \log \sum_{j=1}^{k}P(x^{(i)}, z^{(i)}=j; \theta)\)
By introducing an extra variable \(\displaystyle Q^{(i)}_{(j)}\) we'll be able to apply Jensen's inequality to the concave log function.
Assume \(\displaystyle Q^{(i)}{(j)}\) is a probability mass function.
\(\displaystyle l(\theta) = \sum_{i=1}^{m} \log \sum_{j=1}^{k}P(x^{(i)}, z^{(i)}=j; \theta) \)
\(\displaystyle =\sum_{i=1}^{m} \log \sum_{j=1}^{k} \frac{Q^{(i)}_{(j)}}{Q^{(i)}_{(j)}} P(x^{(i)}, z^{(i)}=j; \theta) \)
\(\displaystyle \geq \sum_{i=1}^{m} \sum_{j=1}^{k} Q^{(i)}_{(j)} \log(\frac{P(x^{(i)}, z^{(i)}=j; \theta) }{Q^{(i)}_{(j)}}) \)
\(\displaystyle \implies \log\left[E_{Q}\left(\frac{Pr(x^{(i)}, z^{(i)}; \theta)}{Q^{(i)}_{(j)}}\right)\right] \geq E_{Q} \left[ \log \left(\frac{Pr(x^{(i)}, q^{(i)}; \theta)}{Q^{(i)}(j)} \right)\right] \)

E-Step

We will fix \(\displaystyle \theta\) and optimize wrt \(\displaystyle Q\).
Jensen's inequality holds with equality iff either the function is linear or if the random variable is degenerate.
We will assume \(\displaystyle \frac{P(x^{(i)}, z^{(i)}=j; \theta) }{Q^{(i)}_{(j)}}\) is a constant.
This implies \(\displaystyle Q^{(i)}(j) = c * P(x^{(i)}, z^{(i)} = j ; \theta)\).
Since Q is a pmf, we have \(\displaystyle Q^{(i)}(j) = \frac{1}{P(x^{(i)})} * P(x^{(i)}, z^{(i)} = j ; \theta) = P(z^{(i)} ; x^{(i)}, \theta)\)

M-Step

We will fix \(\displaystyle Q\) and optimize wrt \(\displaystyle \theta\)