Search

The AI Spring: What Is Machine Learning?

VenusMoon
Jan 20
3 min read

Updated: Feb 4

Until just a few years ago, computers were considered "dumb," merely following pre-recorded rules to generate results. Today, however, they’ve evolved—they’ve started learning as well.

Humans are unique beings on this planet. We post on social media, learn Bharatanatyam (a classical Indian dance form), sing songs, and read posts. We continuously learn and practice a variety of skills. This is what we refer to as human learning.

Similarly, training computers to carry out tasks on their own is known as machine learning. As Arthur Samuel once described it, machine learning is the

"field of study that gives computers the ability to learn without being explicitly programmed."

A computer, when trained through a series of steps, can learn to do things like write a new poem, write code efficiently, draft emails, and much more.

Quick links

#1: What is ML?

#2: How does ML Work?

#3: Example

#1: What Is ML?

Machine learning occurs when a system is trained to perform tasks without needing code for each task, even if the task is not hard-coded or specifically trained for. In essence, the machine is learning. This learning happens when the system is provided with large amounts of data and uses it to find relationships within the data.

For example, if you want a computer to predict how many likes your social media post will attract, you'd collect relevant data, such as previous post impressions, engagements, and replies. These are called input variables or features (x). Your machine learning system will use this data to predict the number of likes your post will get. This predicted number is the output variable (y). The process can be represented as: x → ML → y

#2: How Does ML Work?

How will the machine learning system predict

y = 21 likes

for your next post? It works by learning the relationship between post likes (y) and other features (x).

Let’s look at a simple mathematical relationship:

y = Wx + b

The machine learning system will determine values for W and b, called weights and biases, respectively. It will take the features (x) of your next post and, using the above equation, predict the number of likes (y). This prediction will remain valid for all posts, even if you write a post a year from now. This is the essence of machine learning.

Essentially, you are providing a lot of data (x, y) and training the ML model to find the best values for W and b so that it predicts values that are very close to the actual ones.

What is the primary goal of machine learning?

Learning from data
To manually program all possible scenarios

Check

#3: Example

10-minute grocery delivery apps rely on machine learning to predict the expected time of arrival (ETA) for your grocery bags. They do this by analysing a variety of factors, including:

The distance between your home or neighbourhood and their store,
The type of groceries ordered (e.g., fresh produce vs. non-perishables),
The delivery partner assigned to your order,
Historical delivery data, such as how long it has taken other partners to deliver items to your location,
Real-time factors like traffic conditions, roadblocks, and weather.

By using this data, the machine learning system can calculate an accurate ETA and improve delivery efficiency over time.

In the equation y = Wx + b, what does the "x" represent?