Lecture4 Foundations Supervised Learning

lecture4-foundations-supervised-learning
September 15, 2020
1 Lecture 4: Foundations of Supervised Learning
1.0.1 Applied Machine Learning
Volodymyr KuleshovCornell Tech
2 Why Does Supervised Learning Work?
Prevously, we learned about supervised learning, derived our first algorithm, and used it to predict
diabetes risk.
In this lecture, we are going to dive deeper into why supevised learning really works.
3 Part 1: Data Distribution
First, let’s look at the data, and define where it comes from.
Later, this will be useful to precisely define when supervised learning is guaranteed to work.
4 Review: Components of A Supervised Machine Learning Prob-

lem
At a high level, a supervised machine learning problem has the following structure:
Training Dataset + Learning Algorithm → Predictive Model

| {z } | {z }
Attributes + Features Model Class + Objective + Optimizer
Where does the dataset come from?
1
5 Data Distribution
We will assume that the dataset is sampled from a probability distribution P, which we will call
the data distribution. We will denote this as
x, y ∼ P.
The training set D = {(x(i) , y (i) ) | i = 1, 2, ..., n} consists of independent and identicaly distributed
(IID) samples from P.
6 Data Distribution: IID Sampling
The key assumption in that the training examples are independent and identicaly distributed (IID).
* Each training example is from the same distribution. * This distribution doesn’t depend on
previous training examples.
Example: Flipping a coin. Each flip has same probability of heads & tails and doesn’t depend on
previous flips.
Counter-Example: Yearly census data. The population in each year will be close to that of the
previous year.
7 Data Distribution: Example
Let’s implement an example of a data distribution in numpy.

[1]: import numpy as np
np.random.seed(0)
def true_fn(X):
return np.cos(1.5 * np.pi * X)
Let’s visualize it.

[2]: import matplotlib.pyplot as plt
plt.rcParams['figure.figsize'] = [12, 4]
X_test = np.linspace(0, 1, 100)

plt.plot(X_test, true_fn(X_test), label="True function")
plt.legend()
[2]: <matplotlib.legend.Legend at 0x120e92668>
2
Let’s now draw samples from the distribution. We will generate random x, and then generate
random y using
y = f (x) + ϵ
for a random noise variable ϵ.
[3]: n_samples = 30
X = np.sort(np.random.rand(n_samples))
y = true_fn(X) + np.random.randn(n_samples) * 0.1
We can visualize the samples.

[4]: plt.plot(X_test, true_fn(X_test), label="True function")
plt.scatter(X, y, edgecolor='b', s=20, label="Samples")
plt.legend()
[4]: <matplotlib.legend.Legend at 0x12111c860>
3
8 Data Distribution: Motivation
Why assume that the dataset is sampled from a distribution?

• There is inherent uncertainty in the data. The data may consist of noisy measurements
(readings from an imperfect thermometer).
• There is uncertainty in the process we model. If y is a stock price, there is randomness in the
market that cannot be modeled.
• We can use probability and statistics to analyze supervised learning algorithms and prove
that they work.
# Part 2: Why Does Supervised Learning Work?
We made the assumption that the training dataset is sampled from a data distribution.
Let’s now use it to gain intuition about why supervised learning works.
9 Review: Data Distribution
We will assume that the dataset is sampled from a probability distribution P, which we will call
the data distribution. We will denote this as
x, y ∼ P.
The training set D = {(x(i) , y (i) ) | i = 1, 2, ..., n} consists of independent and identicaly distributed
(IID) samples from P.
10 Review: Supervised Learning Model
We’ll say that a model is a function

f :X →Y
that maps inputs x ∈ X to targets y ∈ Y.
11 What Makes A Good Model?
There are several things we may want out of a good model: 1. Interpretable features that explain
how x affects y. 2. Confidence intervals around y (we will see later how to obtain these) 3. Accurate
predictions of the targets y from inputs x.
In this lecture, we fill focus on the latter.
4
12 Hold-Out Dataset: Definition
A hold-out dataset
Ḋ = {(ẋ(i) , ẏ (i) ) | i = 1, 2, ..., m}
is another dataset that is sampled IID from the same distribution P as the training dataset D and
the two datasets are disjoint.
Let’s genenerate a hold-out dataset for the example we saw earlier.
[5]: import numpy as np
import matplotlib.pyplot as plt
np.random.seed(0)
def true_fn(X):
return np.cos(1.5 * np.pi * X)

plt.legend()
[5]: <matplotlib.legend.Legend at 0x12116be48>
Let’s genenerate a hold-out dataset for the example we saw earlier.

[6]: n_samples, n_holdout_samples = 30, 30
X_holdout = np.sort(np.random.rand(n_holdout_samples))
y_holdout = true_fn(X_holdout) + np.random.randn(n_holdout_samples) * 0.1

plt.scatter(X_holdout, y_holdout, edgecolor='r', s=20, label="Holdout Samples")
5
plt.legend()
[6]: <matplotlib.legend.Legend at 0x121440f28>
13 Defining What is an Accurate Model
Suppose that we have a function isaccurate(y, y ′ ) that determines if y is an accurate estimate of

y ′ , e.g.: * Is the the target variable close enough to the true target?
isaccurate(y, y ′ ) = true if (|y − y ′ | is small), else false
• Did we predict the right class?
isaccurate(y, y ′ ) = true if (y = y ′ ) else false
This defines accuracy on a data point. We say a supervised learning model is accurate if it correctly
predicts the target on new (held-out) data.
We can say that a predictive model f is accurate if it’s probability of making an error on a random
holdout sample is small:
1 − P [isaccurate(ẏ, f (ẋ))] ≤ ϵ
for ẋ, ẏ ∼ P, for some small ϵ > 0 and some definition of accuracy.
We can also say that a predictive model f is inaccurate if it’s probability of making an error on a
random holdout sample is large:
1 − P [isaccurate(ẏ, f (ẋ))] ≥ ϵ
or equivalently
P [isaccurate(ẏ, f (ẋ))] ≤ 1 − ϵ.
6
14 Generalization
In machine learning, generalization is the property of predictive models to achieve good perfor-
mance on new, heldout data that is distinct from the training set.
Will supervised learning return a model that generalizes?
15 Recall: Supervised Learning
Recall our intuitive definition of supervised learning.

1. First, we collect a dataset of labeled training examples.
2. We train a model to output accurate predictions on this dataset.
3. When the model sees new, similar data, it will also be accurate.
16 Applying Supervised Learning
Let’s say that we apply our supervised learning to our problem. 1. We define a model class M
containing H different models. 2. One of these models fits the training data perfectly (is accurate
on every point) and we choose that model.
(These are assumptions that could be changed.)
17 Why Supervised Learning Works
Claim: The probability that supervised learning will return an inaccurate model decreases expo-
nentially with training set size n.
1. A model f is inaccurate if P [isaccurate(ẏ, f (ẋ))]
∏n≤ 1−ϵ.
[ The probability that an] inaccurate
model f perfectly fits the training set is at most i=1 P isaccurate(y , f (x(i) )) ≤ (1−ϵ)n .
(i)
2. We have H models in M, and any of them could be in accurate. The probability that at
least one the at most H inaccurate models willl fit the training set perfectly is ≤ H(1 − ϵ)n .
Therefore, the claim holds.
# Part 3: Overfitting and Underfitting
Let’s now dive deeper into the concept of generalization and two possible failure modes of supervised
learning: overfitting and underfitting.
18 Review: Generalization
We will assume that the dataset is governed by a probability distribution P, which we will call the
data distribution. We will denote this as
x, y ∼ P.
7
˙ , y (i)
A hold-out set Ḋ = {(x(i) ˙ ) | i = 1, 2, ..., n} consists of independent and identicaly distributed
(IID) samples from P and is distinct from the training set.
A model that generalizes is accurate on a hold-out set.
19 Review: Polynomial Regression
In 1D polynomial regression, we fit a model
fθ (x) := θ⊤ ϕ(x)
that is linear in θ but non-linear in x because the features ϕ(x) : R → Rp are non-linear.
By using polynomial features such as ϕ(x) = [1 x . . . xp ], we can fit any polynomial of degree p.
20 Polynomials Better Fit the Data
When we switch from linear models to polynomials, we can better fit the data and increase the
accuracy of our models.
Consider the synthetic dataset that we have seen earlier.
[7]: from sklearn.pipeline import Pipeline
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression
np.random.seed(0)
n_samples = 30

[7]: <matplotlib.collections.PathCollection at 0x12e0c58d0>
8
Although fitting a linear model does not work well, qudratic or cubic polynomials improve the fit.
[8]: degrees = [1, 2, 3]
plt.figure(figsize=(14, 5))
for i in range(len(degrees)):
ax = plt.subplot(1, len(degrees), i + 1)
polynomial_features = PolynomialFeatures(degree=degrees[i],␣
,→include_bias=False)
linear_regression = LinearRegression()
pipeline = Pipeline([("pf", polynomial_features), ("lr",␣
,→linear_regression)])
pipeline.fit(X[:, np.newaxis], y)
ax.plot(X_test, true_fn(X_test), label="True function")

ax.plot(X_test, pipeline.predict(X_test[:, np.newaxis]), label="Model")
ax.scatter(X, y, edgecolor='b', s=20, label="Samples")
ax.set_xlim((0, 1))
ax.set_ylim((-2, 2))
ax.legend(loc="best")
ax.set_title("Polynomial of Degree {}".format(degrees[i]))
21 Towards Higher-Degree Polynomial Features?
As we increase the complexity of our model class M to even higher degree polynomials, we are able
to fit the data increasingly even better.
What happens if we further increase the degree of the polynomial?
9
[10]: degrees = [30]

ax.set_xlim((0, 1))
22 The Problem With Increasing Model Capacity
As the degree of the polynomial increases to the size of the dataset, we are increasingly able to fit
every point in the dataset.
However, this results in a highly irregular curve: its behavior outside the training set is wildly
inaccurate.
10
23 Overfitting
Overfitting is one of the most common failure modes of machine learning. * A very expressive
model (a high degree polynomial) fits the training dataset perfectly. * The model also makes
wildly incorrect prediction outside this dataset, and doesn’t generalize.
24 Underfitting
A related failure mode is underfitting.

• A small model (e.g. a straight line), will not fit the training data well.
• Held-out data is similar to training data, so it will not be accurate either.
Finding the tradeoff between overfitting and underfitting is one of the main challenges in applying
machine learning.
25 Overfitting vs. Underfitting: Evaluation
We can measure overfitting and underfitting by estimating accuracy on held-out data and comparing
it to the training data. * If training perforance is high but held-out performance is low, we are
overfitting. * If training perforance is low but held-out performance is low, we are underfitting.
[11]: degrees = [1, 20, 5]
titles = ['Underfitting', 'Overfitting', 'A Good Fit']

ax.scatter(X, y, edgecolor='b', s=20, label="Samples", alpha=0.2)
ax.scatter(X_holdout[::3], y_holdout[::3], edgecolor='r', s=20,␣
,→label="Samples")
ax.set_xlim((0, 1))
ax.set_title("{} (Degree {})".format(titles[i], degrees[i]))
11
ax.text(0.05,-1.7, 'Holdout MSE: %.4f' % ((y_holdout-pipeline.
predict(X_holdout[:, np.newaxis]))**2).mean())
,→
26 Dealing with Underfitting
Balancing overfitting vs. underfitting is a major challenges in applying machine learning. Briefly,
here are some approaches: * To fight under-fitting, we may increase our model class to encompass
more expressive models. * We may also create richer features for the data that will make the
dataset easier to fit.
27 Dealing with Overfitting
We will see many ways of dealing with overftting, but here are some ideas: * If we’re overfitting,
we may reduce the complexity of our model by reducing the size of M * We may also modify our
objective to penalize complex models that may overfit the data.
# Part 4: Regularization
We will now see a very important way to reduce overfitting — regularization. We will also see
several important new algorithms.
28 Review: Generalization
We will assume that the dataset is governed by a probability distribution P, which we will call the
data distribution. We will denote this as
x, y ∼ P.
˙ , y (i)
A hold-out set Ḋ = {(x(i) ˙ ) | i = 1, 2, ..., n} consists of independent and identicaly distributed
(IID) samples from P and is distinct from the training set.
12
29 Review: Overfitting
Overfitting is one of the most common failure modes of machine learning. * A very expressive
model (a high degree polynomial) fits the training dataset perfectly. * The model also makes
wildly incorrect prediction outside this dataset, and doesn’t generalize.
We can visualize overfitting by trying to fit a small dataset with a high degree polynomial.
[12]: degrees = [30]

ax.set_xlim((0, 1))
13
30 Regularization: Intuition
The idea of regularization is to penalize complex models that may overfit the data.
In the previous example, a less complex would rely less on polynomial terms of high degree.
31 Regularization: Definition
The idea of regularization is to train models with an augmented objective J : M → R defined over
a training dataset D of size n as
1∑
n
J(f ) = L(y (i) , f (x(i) )) + λ · R(f ).
n
i=1
Let’s dissect the components of this objective:
1∑
n
J(f ) = L(y (i) , f (x(i) )) + λ · R(f ).
n
i=1
• A loss function L(y, f (x)) such as the mean squared error.

• A regularizer R : M → R that penalizes models that are overly complex.
• A regularization parameter λ > 0, which controls the strength of the regularizer.
When the model fθ is parametrized by parameters θ, we can also use the following notation:
1∑
n
J(θ) = L(y (i) , fθ (x(i) )) + λ · R(θ).
n
i=1
32 L2 Regularization: Definition
How can we define a regularizer R : M → R to control the complexity of a model f ∈ M?

In the context of linear models f (x) = θ⊤ x, a widely used approach is L2 regularization, which
defines the following objective:
1∑
n
λ
J(θ) = L(y (i) , θ⊤ x(i) ) + · ||θ||22 .
n 2
i=1
Let’s dissect the components of this objective.
1∑
n
λ
J(θ) = L(y (i) , θ⊤ x(i) ) + · ||θ||22 .
n 2
i=1
∑d
• The regularizer R : M → R is the function R(θ) = ||θ||22 = 2
j=1 θj . This is also known as
the L2 norm of θ.
14
• The regularizer penalizes large parameters. This prevents us from over-relying on any single
feature and penalizes wildly irregular solutions.
• L2 regularization can be used with most models (linear, neural, etc.)
33 L2 Regularization for Polynomial Regression
Let’s consider an application to the polynomial model we have seen so far. Given polynomial
features ϕ(x), we optimize the following objective:
1 ∑ ( (i) )2 λ
n
J(θ) = y − θ⊤ ϕ(x(i) ) + · ||θ||22 .
2n 2
i=1
We are going to implement regularized and standard polynomial regression on three random training
sets sampled from the same distribution.
[13]: from sklearn.linear_model import Ridge
degrees = [15, 15, 15]

for idx, i in enumerate(range(len(degrees))):
# sample a dataset
np.random.seed(idx)
n_samples = 30
# fit a least squares model

# fit a Ridge model

linear_regression = Ridge(alpha=0.1) # sklearn uses alpha instead of lambda

pipeline2 = Pipeline([("pf", polynomial_features), ("lr",␣
pipeline2.fit(X[:, np.newaxis], y)
# visualize results
# ax.plot(X_test, true_fn(X_test), label="True function")
15
ax.plot(X_test, pipeline.predict(X_test[:, np.newaxis]), label="No␣
,→Regularization")
ax.plot(X_test, pipeline2.predict(X_test[:, np.newaxis]), label="L2␣
,→Regularization")

ax.set_xlim((0, 1))
ax.set_title("Dataset sample #{}".format(idx))
We can show that by usinng small weights, we prevent the model from learning irregular functions.
[14]: print('Non-regularized weights of the polynomial model need to be large to fit␣
,→every point:')
print(pipeline.named_steps['lr'].coef_[:4])
print()
print('By regularizing the weights to be small, we force the curve to be more␣

,→regular:')
print(pipeline2.named_steps['lr'].coef_[:4])
Non-regularized weights of the polynomial model need to be large to fit every

point:
[-3.02370887e+03 1.16528860e+05 -2.44724185e+06 3.20288837e+07]
By regularizing the weights to be small, we force the curve to be more regular:

[-2.70114811 -1.20575056 -0.09210716 0.44301292]
16
34 How to Choose λ?
In brief, the most common approach is to choose the value of λ that results in the best performance
on a held-out validation set.
We will later see this strategies and several other in more detail
35 Normal Equations for Regularized Models
How, do we fit regularized models? In the linear case, we can do this easily by deriving generalized
normal equations!
Let L(θ) = 12 (Xθ − y)⊤ (Xθ − y) be our least squares objective. We can write the Ridge objective
as:
1 1
J(θ) = (Xθ − y)⊤ (Xθ − y) + λ||θ||22
2 2
This allows us to derive the gradient as follows:
( )
1 ⊤ 1
∇θ J(θ) = ∇θ (Xθ − y) (Xθ − y) + λ||θ||2
2
2 2
( )
1
= ∇θ L(θ) + λ||θ||22
2
= ∇θ L(θ) + λθ
= (X ⊤ X)θ − X ⊤ y + λθ
= (X ⊤ X + λI)θ − X ⊤ y
We used the derivation of the normal equations for least squares to obtain ∇θ L(θ) as well as the
fact that: ∇x x⊤ x = 2x.
We can set the gradient to zero to obtain normal equations for the Ridge model:
(X ⊤ X + λI)θ = X ⊤ y.
Hence, the value θ∗ that minimizes this objective is given by:
θ∗ = (X ⊤ X + λI)−1 X ⊤ y.
Note that the matrix (X ⊤ X +λI) is always invertible, which addresses a problem with least squares
that we saw earlier.
36 Algorithm: Ridge Regression
• Type: Supervised learning (regression)

• Model family: Linear models
• Objective function: L2-regularized mean squared error
17
• Optimizer: Normal equations
# Part 5: Regularization and Sparsity
We will now look another form of regularization, which will have an important new property called
sparsity.
37 Regularization: Definition
The idea of regularization is to train models with an augmented objective J : M → R defined over
a training dataset D of size n as
1∑
n
J(f ) = L(y (i) , f (x(i) )) + λ · R(f ).
n
i=1
Let’s dissect the components of this objective:
1∑
n
J(f ) = L(y (i) , f (x(i) )) + λ · R(f ).
n
i=1
• A loss function L(y, f (x)) such as the mean squared error.

• A regularizer R : M → R that penalizes models that are overly complex.
38 L1 Regularizion: Definition
Another closely related approach to regularization is to penalize the size of the weights using the
L1 norm.
In the context of linear models f (x) = θ⊤ x, L1 regularization yields the following objective:
1∑
n
J(θ) = L(y (i) , θ⊤ x(i) ) + λ · ||θ||1 .
n
i=1
Let’s dissect the components of this objective.
1∑
n
J(θ) = L(y (i) , θ⊤ x(i) ) + λ · ||θ||1 .
n
i=1
∑d
• The regularizer R : M → R is the function R(θ) = ||θ||1 = j=1 |θj |. This is also known as
the L1 norm of θ.
• The regularizer also penalizes large weights. It also forces more weights to decay to zero, as
opposed to just being small.
18
39 Algorithm: Lasso
L1-regularized linear regression is also known as the Lasso (least absolute shrinkage and selection
operator).
• Type: Supervised learning (regression)
• Model family: Linear models
• Objective function: L1-regularized mean squared error
• Optimizer: gradient descent, coordinate descent, least angle regression (LARS) and others
40 Regularizing via Constraints
Consider regularized problem with a penalty term:
min L(θ) + λ · R(θ).

θ∈Θ
We may also enforce an explicit constraint on the complexity of the model:
min L(θ)
θ∈Θ
such that R(θ) ≤ λ′
We will not prove this, but solving this problem is equivalent so solving the penalized problem for
some λ > 0 that’s different from λ′ .
In other words, * We can regularize by explicitly enforcing R(θ) to be less than a value instead of
penalizing it. * For each value of λ, we are implicitly setting a constraint of R(θ).
41 Regularizing via Constraints: Example
This is what it looks like for a linear model:
1 ∑ ( (i) )2
n
min y − θ⊤ x(i)
θ∈Θ 2n
i=1
such that ||θ|| ≤ λ′
where || · || can either be the L1 or L2 norm.
42 L1 vs. L2 Regularization
The following image by Divakar Kapil and Hastie et al. explains the difference between the two
norms.
19
43 Sparsity: Definition
A vector is said to be sparse if a large fraction of its entires is zero.

L1-regularized linear regression produces sparse weights. * This is makes the model more inter-
pretable * It also makes it computationally more tractable in very large dimensions.
44 Sparsity: Ridge Model
To better understand sparsity, we will fit L2-regularized linear models to the UCI diabetes dataset
and observe the magnitude of each weight (colored lines) as a function of the regularization param-
eter.
[15]: # based on https://scikit-learn.org/stable/auto_examples/linear_model/
,→plot_ridge_path.html
from sklearn.datasets import load_diabetes

from sklearn.linear_model import Ridge
from matplotlib import pyplot as plt
X, y = load_diabetes(return_X_y=True)
# create ridge coefficients

alphas = np.logspace(-5, 2, )
ridge_coefs = []
for a in alphas:
ridge = Ridge(alpha=a, fit_intercept=False)
ridge.fit(X, y)
ridge_coefs.append(ridge.coef_)
# plot ridge coefficients

plt.plot(alphas, ridge_coefs)
plt.xscale('log')
plt.xlabel('Regularization parameter (lambda)')
plt.ylabel('Magnitude of model parameters')
plt.title('Ridge coefficients as a function of the regularization')
plt.axis('tight')
[15]: (4.466835921509635e-06,
223.872113856834,
-868.4051623855127,
828.0533448059361)
20
45 Sparsity: Lasso Model
The above Ridge model did not produce sparse weights. Let’s now compare it to a Lasso model.
[16]: # Based on: https://scikit-learn.org/stable/auto_examples/linear_model/
,→plot_lasso_lars.html
import warnings
warnings.filterwarnings("ignore")
from sklearn.datasets import load_diabetes
from sklearn.linear_model import lars_path
# create lasso coefficients

_, _, lasso_coefs = lars_path(X, y, method='lasso')
xx = np.sum(np.abs(lasso_coefs.T), axis=1)
# plot ridge coefficients

plt.subplot('121')
plt.plot(alphas, ridge_coefs)
plt.xscale('log')
plt.ylabel('Regularization Strength (alpha)')
plt.ylabel('Coefficents')
plt.title('Ridge coefficients as a function of the regularization')
plt.axis('tight')
# plot lasso coefficients

plt.subplot('122')
plt.plot(3500-xx, lasso_coefs.T)
ymin, ymax = plt.ylim()
plt.xlim(ax.get_xlim()[::-1]) # reverse axis
21
plt.ylabel('Coefficients')
plt.ylabel('Regularization Strength')
plt.title('LASSO Path')
plt.axis('tight')
[16]: (3673.0002477572816,
-133.00520290291772,
-869.3573357636973,
828.4524952229636)
22

Lecture4 Foundations Supervised Learning

Uploaded by

Copyright:

Available Formats

Lecture4 Foundations Supervised Learning

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture4 Foundations Supervised Learning

Uploaded by

Copyright:

Available Formats

lecture4-foundations-supervised-learning

September 15, 2020

1 Lecture 4: Foundations of Supervised Learning

1.0.1 Applied Machine Learning

Volodymyr KuleshovCornell Tech

2 Why Does Supervised Learning Work?

3 Part 1: Data Distribution

4 Review: Components of A Supervised Machine Learning Prob-

Training Dataset + Learning Algorithm → Predictive Model

Where does the dataset come from?

6 Data Distribution: IID Sampling

7 Data Distribution: Example

Let’s implement an example of a data distribution in numpy.

Let’s visualize it.

X_test = np.linspace(0, 1, 100)

[2]: <matplotlib.legend.Legend at 0x120e92668>

We can visualize the samples.

[4]: <matplotlib.legend.Legend at 0x12111c860>

Why assume that the dataset is sampled from a distribution?

9 Review: Data Distribution

10 Review: Supervised Learning Model

We’ll say that a model is a function

11 What Makes A Good Model?

X_test = np.linspace(0, 1, 100)

[5]: <matplotlib.legend.Legend at 0x12116be48>

Let’s genenerate a hold-out dataset for the example we saw earlier.

plt.plot(X_test, true_fn(X_test), label="True function")

[6]: <matplotlib.legend.Legend at 0x121440f28>

13 Defining What is an Accurate Model

Suppose that we have a function isaccurate(y, y ′ ) that determines if y is an accurate estimate of

isaccurate(y, y ′ ) = true if (|y − y ′ | is small), else false

• Did we predict the right class?

isaccurate(y, y ′ ) = true if (y = y ′ ) else false

15 Recall: Supervised Learning

Recall our intuitive definition of supervised learning.

16 Applying Supervised Learning

17 Why Supervised Learning Works

19 Review: Polynomial Regression

In 1D polynomial regression, we fit a model

20 Polynomials Better Fit the Data

X_test = np.linspace(0, 1, 100)

[7]: <matplotlib.collections.PathCollection at 0x12e0c58d0>

ax.plot(X_test, true_fn(X_test), label="True function")

21 Towards Higher-Degree Polynomial Features?

X_test = np.linspace(0, 1, 100)

22 The Problem With Increasing Model Capacity

A related failure mode is underfitting.

25 Overfitting vs. Underfitting: Evaluation

ax.plot(X_test, true_fn(X_test), label="True function")

26 Dealing with Underfitting

27 Dealing with Overfitting

X_test = np.linspace(0, 1, 100)

Let’s dissect the components of this objective:

• A loss function L(y, f (x)) such as the mean squared error.

How can we define a regularizer R : M → R to control the complexity of a model f ∈ M?

Let’s dissect the components of this objective.

33 L2 Regularization for Polynomial Regression

degrees = [15, 15, 15]