Lecture 01. Introduction & Basic Concepts

Lecture video link: https://youtu.be/jGwO_UgTS7I

What is machine learning?

Arthur Samuel (1959). Machine learning: Field of study that gives computers the ability to learn without being explicitly programmed.

Tom Mitchell (1998). Well-posed learning problem: A computer program is said to learn from experience $E$ with respect to some task $T$ and some performance measure $P$, if its performance on $T$, as measured by $P$, improves with experience $E$.

(Andrew Ng asked his friend Tom if he defined the problem to rhyme, but the answer was no… 😊)

Supervised learning

Goal: to learn a mapping from $X$ to $y$ given a dataset of inputs and labels, $(X, y)$.

What Andrew Ng find is …

The most skilled machine learning practitioners are very “strategic.”

You need to make decision when you work on a ML project - collect more data, try different learning algorithms, rent faster GPUs to train your learning algorithm longer, …

→ Become a systematic engineer.

Cocktail party problem

Suppose that you have multiple microphones in a noisy room and record overlapping voices. How can you have an algorithm that separate out the people’s voices so that you get clean recordings of just one voice at a time?

~ unsupervised problem