Machine Learning: Introduction
From a beginner’s perspective
The ability to learn is a skill;
The willingness to learn is a choice.
- Brain Herbert.
Definition
Machine Learning (ML) is a branch of Artificial Intelligence (AI) that uses algorithms and data for finding insights, discovering patterns, like the human brain. The term Machine Learning consists of two words:
Machine-like computer, mobile phone, or any other device, and
Learning -the ability to learn.
The term Machine Learning was coined by Arthur Samuel in 1959. He defined machine learning as a “Field of study that gives computers the ability to learn without being explicitly programmed.”
Applications of Machine Learning
There are limitless applications of Machine Learning available in the real world. If we observe closely, we will find many applications based on the Machine Learning algorithms:
1. E-mails Filteration: This is one of the most famous applications of ML. If you are a Gmail user, you must have noticed that your mails have been automatically categorized as Spam, Social, Promotional, etc. This is done by using ML classification algorithms.
2. Image Recognition: This application of ML is also widely used. Nowadays, mobile phones have facial recognition screen locks. Also, Cameras automatically detect our face and encloses it with a circle or square. We also use this application to identify objects, places, etc. The technology behind this is ML’s Image detection and Facial recognition algorithm.
3. Speech Recognition: We all are familiar with the Speech-to-Text function in our devices. Google Search also has an option to search by voice. Also, many voice assistants such as Alexa, Siri, Google Assistant use ML’s Speech Recognition application.
4. Recommendation Engines: Netflix, YouTube, Amazon, etc., companies use ML Recommendation Engines for products or videos recommendations to their users.
5. Medical Diagnosis: Machine Learning has achieved a lot in the medical field in very little time. ML application is used to diagnose diseases like Cancer, Brain Tumors, etc.
6. Stock Market Trading: We can predict trends in Stocks using ML application.
7. Fake News Detection: We use NLP to classify whether or not news is fake.
8. Fraud Detection: This ML application makes our online transactions safe and secure. Neural Network helps us by checking if it is a genuine transaction or a fraud transaction.
There are many more such applications available in the real world.
Is Machine Learning the same as Automation?
There is always a debate that Machine Learning is the same as Automation. But, this is not true. Automation is a set of instructions build to perform over a fixed format of data repetitively. Automation needs to be maintained, redefined, and updated from time to time. For discovering new patterns, analyzing new insights would be exhausting as well as expensive over time.
This is why we need Machine Learning. We train models on ML algorithms over different sets of data. This lets them learn efficiently, extract patterns, and figure things out themselves.
Types of Machine Learning
The main task of Machine Learning is to explore and construct an algorithm that can learn from the historical data and make predictions on the new input data. Then, we need to introduce an evaluation function called loss or cost function, which measures how well our model is learning.
Based on historical learning data, machine learning tasks are classified into three categories:
- Supervised Learning: Supervised Learning is a somewhat similar way of learning for toddlers. We help them to distinguish among animals until they start identifying them on their own.
In Supervised Machine Learning, the learning data come up with labels. In layman terms, we present input and output data to the machine, and the goal is to find a rule that maps input to output. This training process continues until the machine achieves desired accuracy. It is used in Regression and Classification problems.
Some of the used applications of Supervised Learning are House Price Prediction, Product or Movie Recommendation, Sales Price Forecasting, Spam E-mails Classification, Image Classification. - Unsupervised Learning: In Unsupervised Machine Learning, the learning data are unlabeled. The goal of using Unsupervised Learning is to discover hidden patterns and information in the data.
Some of the used applications of Unsupervised Learning are Anomaly Detection, Recommendation Engines, Clustering of Articles. - Reinforcement Learning: In Reinforcement Learning, the program provides feedback so that machine adapts to dynamic conditions to achieve the goal. The feedbacks are generally rewards and punishments. The machine evaluates its performance based on feedback.
Reinforcement Learning is one of the hottest topics in the field of Artificial Intelligence these days. Some of the used applications of Reinforcement Learning are Self Driving Cars, Robots, Computer Games.
EndNotes
Here I conclude the topic. This was some basic information about Machine Learning truly from a Beginners's perspective. Thanks for being with me till here. This is my first article, and I hope you got to learn something from this article. I’ll be posting more articles on ML in recent times.
References:
- Book: Python Machine Learning By Example — Yuxi (Hayden) Liu
- Wikipedia
- Geeks for Geeks
