Introduction to Machine Learning

Study Notes

Machine learning is a subset of artificial intelligence (AI) that enables systems to learn and improve automatically through experience without explicit programming.
The term "machine learning" was coined by Arthur Samuel in 1959, defining it as a field of study that allows computers to learn from data.

The machine learning process improves models over time using training data, allowing for predictions based on evolved models.
A machine learning model mathematically represents a real-world process, mapping inputs to outputs through a function.

Key to machine learning is the extraction of patterns from historical data to apply to new data.
Data quality significantly affects the accuracy and effectiveness of machine learning outputs.

Supervised Learning: Trains on labeled data to predict outputs for unseen data using:
- Classification: Predicts discrete labels (e.g., classify images as "cat" or "dog").
- Regression: Predicts continuous outcomes (e.g., estimate house prices).
Unsupervised Learning: Identifies patterns in data without labeled outputs, primarily using clustering algorithms to group similar data points.
Semi-Supervised Learning: Combines a small amount of labeled data with a large pool of unlabeled data. It typically involves:
- Initial model training with labeled data
- Predicting labels for unlabeled data
- Creating a new model with both labeled and newly labeled data.
Reinforcement Learning: Utilizes trial and error to discover the relationship between goals and sequences of actions to achieve success. Applications include:
- Robotics navigating environments
- Self-driving vehicles

Supervised Learning algorithms include:
- Decision trees
- Neural networks and deep learning
Unsupervised Learning algorithms include:
- K-means clustering for grouping data points based on similarity.