Machine Learning Introduction

Arthur Samuel (1959)

“Machine Learning is the field of study that gives computers the ability to learn without being explicitly programmed.”

What is Machine Learning?

Machine Learning is a subset of artificial intelligence that enables systems to learn and improve from experience. Instead of explicit programming, ML algorithms build mathematical models based on training data.

Types of Machine Learning

graph TD
    A[Machine Learning] --> B[Supervised Learning]
    A --> C[Unsupervised Learning]
    A --> D[Reinforcement Learning]
    B --> E[Classification]
    B --> F[Regression]
    C --> G[Clustering]
    C --> H[Dimensionality Reduction]
    D --> I[Policy Learning]
    D --> J[Q-Learning]

Supervised Learning

In supervised learning, we have labeled data:

$y=f(x)+ϵ$

Where:

• $y$ is the target variable
• $x$ represents features
• $ϵ$ is noise

Linear Regression

The simplest model: $\hat{y}={\beta }_0+{\beta }_1{x}_1+{\beta }_2{x}_2+\dots +{\beta }_n{x}_n$

from sklearn.linear_model import LinearRegression
import numpy as np
 
# Sample data
X = np.array([[1], [2], [3], [4], [5]])
y = np.array([2, 4, 5, 4, 5])
 
# Train model
model = LinearRegression()
model.fit(X, y)
 
# Predict
print(f"Coefficient: {model.coef_[0]:.2f}")
print(f"Intercept: {model.intercept_:.2f}")

Unsupervised Learning

No labels provided — the algorithm finds patterns:

from sklearn.cluster import KMeans
 
kmeans = KMeans(n_clusters=3, random_state=42)
clusters = kmeans.fit_predict(data)

K-Means Algorithm

1. Initialize $k$ centroids randomly
2. Assign each point to nearest centroid
3. Recalculate centroids as mean of assigned points
4. Repeat until convergence

Neural Networks

Deep Learning

Neural networks with multiple hidden layers are called “deep” neural networks. They form the basis of modern AI applications.

The forward pass of a neural network:

$z^{[l]}=W^{[l]}{a}^{[l-1]}+b^{[l]}$ $a^{[l]}=g(z^{[l]})$

Where $g$ is an activation function like ReLU:

$g(z)=\max (0,z)$

Evaluation Metrics

Metric	Use Case	Formula
Accuracy	Classification	$\frac{TP+TN}{TP+TN+FP+FN}$
Precision	When FP is costly	$\frac{TP}{TP+FP}$
Recall	When FN is costly	$\frac{TP}{TP+FN}$
F1 Score	Balanced measure	$2\cdot \frac{Precision\cdot Recall}{Precision+Recall}$

Common Mistake

Don’t confuse correlation with causation! A model might find spurious patterns in training data that don’t generalize. See Book Notes - Thinking Fast and Slow for cognitive biases that affect data interpretation.

Learning Path

My ML Journey

• Python basics — see Web Development Basics
• Linear regression
• Classification algorithms
• Deep learning with PyTorch
• Natural Language Processing
• Computer Vision

Related Notes

• How I Take Notes — Organizing ML study materials
• Git Version Control — Version controlling ML experiments
• Docker Containerization — Reproducible ML environments

---

*Tags: machine-learning ai python data-science