What Is Machine Learning?

provides machines
Machine learning (ML) isa discipline of artificial intelligence (AI) that
identifying
with the ability toautomatically learn from data and past experiences while
patternsto make predictions with minimal human intervention.
Machine learning methods enable computers to operate autonomously without explicit
programming. ML applications are fed with new data, and they can independently learn, grow,
develop, and adapt.
Machine learning derives insightful information from large volumes of data by leveraging
algorithms to identify patterns and learn in an iterative process. ML algorithms use computation
methods to learndirectly from data instead of relying on any predetermined equation that may
serve as a model.

The performance of ML algorithms adaptively improves with an increase in the number of

available samples during the learning' processes. For example, deep learning is asub-domain
of machine learning that trains computers to imitate natural human traits like learning from
examples. It offers better performance parameters than conventional ML algorithms.
While machine learning is not a new concept - dating back to World War ll when the Enigma
Machine was used - the ability to apply complex mathematical calculations automatically to
growing volumes and varieties of available data isa relatively recent development.

Today, with the rise of big data, loT, and ubiquitous computing, machine learning has become
essential for solving problems across numerous areas, Such as

Computational finance (credit scoring, algorithmic trading)

Computer vision (facial recognition, motion tracking, object detection)
Computational biology (DNA sequencing, brain tumor detection, drug discovery)
maintenance)
Automotive, aerospace, and manufacturing (predictive

Natural language processing (voice recognition)

How does machine learning workK?
learning algorithms are molded on a training dataset to create a model. As new input
Machine
introduced to the trained ML algorithm, it uses the developed model to make a
data is
prediction.
 HOW DOES MACHINE L
LEARNING WORK?
Training
data

(Unacceptable) (Acceptable)

Train ML
algorithm Accuracy Successful
model

Model input
data

New input
data ML algorithm Prediction
How Machine Learning Works
Note: The above illustration discloses a
high-level use case
learning examples may involve many other factors, variables,scenario. However, typical machine
and steps.
Further, the prediction is checked for accuracy. Based on its accuracy, the
ML
deployed or trained repeatedly with an augmented training dataset until the algorithm is either
achieved. desired accuracy is

See More: What Is Artificial Intelligence (Al) as a Service?

Trends
Definition, Architecture, and

Types of Machine Learning

Machine learning algorithms can be trained in manyways, with each method having its pros ana
cons. Based on these methods and ways of learning, machine learning is broadly categorizea
into four main types:
 TYPES OF
MACHINE LEARNING

GD

Supervised Unsupervised
Machine Learning Machine Learning Semi-Supervised
Learning
Reinforcement
Learning

Types of Machine Learning

1.Supervised machine learning
This type of ML involves supervision, where machines are trained on labeled datasets and
enabled to predict outputs based on the provided training. The labeled dataset specifies that
is trained with the
some input and output parameters are already mapped. Hence, the machine
input and corresponding output. A device is made to predict the outcome using thetest dataset.
in subsequent phases.
and crow images. Initially, the machine is
For example, consider an inputdataset of parrot
parrot and crow's color, eyes, shape, and size.
trained to understand the pictures, including the
the machine is expected to identify
Post-training,an input picture of a parrot is provided, and features of the
object and predict the output. The trained machine checks for the various
the prediction. This is
such as color, eyes, shape, etc., in the input picture, to make a final
object,
in supervised machine learning.
the process of object identification
variable (a) with
objective of the supervised learning technique is to map the input
The primary broad
Supervised machine learning is further classified into two
the output variable (b).
categories:
algorithms that address classification problems where
Classification: These refer to no, true or false, male or female,
categorical: for example, yes or email
the output variable is
category are evident in spam detection and
applications of this
etc. Real-world
filtering. Forest Algorithm, Decision Tree
includethe Random
Some known classification algorithms Support Vector Machine Algorithm.
Algorithm, and
Algorithm, Logistic Regression
regression problems where input and output
algorithms handle continuous output
Kegression: Regression known to predict
relationship. These are etc.
variables have alinear prediction, markettrend analysis,
weather
variables. Examples include
Popular regression
Regression Algorithm,algorithms
Decision
include the Simple Linear
Tree Agorithm, and Regression Algorithm,
2.
Unsupervised machine learning Lasso Regression. Multivariate
Unsupervised learning refers to a learning technique that's
machine is trained using an unlabeled devoid of supervision. Here, the
SupervisSion. An dataset and is enabled
the input's unsupervised learning algorithm aims to grouptothepredict the output without any
similarities, differences, and patterns. unsorted dataset based on
For example, consider an input
dataset of images of a fruit-filled
not known to the machine
learning model. When we input container. Here, the images are
task of the model is to identify the the dataset into the ML
pattern of objects, such as color, shape, or model, the
inthe input images and categorize
them. Upon categorization, the machine then differences seen
Output as it gets tested with a test predicts the
dataset.
Unsupervised machine learning is further classified into two types:
Clustering: The clustering technique refers to grouping objects into clusters based on
parameters such as similarities or differences between objects. For example, grouping
customers by the products they purchase.
Some known clustering algorithms include the K-Means Clustering Algorithm, Mean-Shift
Algorithm, DBSCAN Algorithm, Principal Component Analysis, and Independent Component
Analysis.
Association: Association learning refers to identifying typical relations between the
variables of a large dataset. It determines the dependency of various data items and
maps associated variables. Typical applications include web usage mining and market
data analysis.
Popular algorithms obeying association rules include the AprioriAlgorithm, Eclat Algorithm, and
FP-Growth Algorithm.
3.Semi-supervised learning
Semi-supervised learning comprises characteristics of both supervised and unsupervised
to train its
machine learning. It uses the combination of labeled and unlabeled datasets
the drawbacks
algorithms. Using both types of datasets, semi-supervised learning overcomes
of the options mentioned above.
concept under a teacher's
Consider an example of a college student. Astudent learning a
learning, a student self
supervision in college is termed supervised learning. In unsupervised
guidance. Meanwhile, a student revising
learns the same concept at home without ateacher's
teacher in college is a semi-supervised form
the concept after learning under the direction of a
of learning.
4. Reinforcement learning
the Al component automaticaly
Reinforcement learning is a feedback-based process. Here,
method, takes action, learns from experienceS,
takes stock of its surroundings by the hit &trial penalized Tor
improves performance. The component is rewarded for each good action and
and
reinforcement learning component aims to maximize the rewards
every wrong move. Thus, the
by performing good actions.
Unlike Supervised learning,
experiences reinforcement
only. Consider video games.
learning lacks labeled data, and the
agents learn via
move of the Here, the game specifies the
reinforcemnent agent defines its state. The agent is environment, and each
punishment and entitled to
rewards, thereby affecting the overall game score. The receive feedback via
agent is to achieve a high score. ultimate goal of the
Reinforcement learning is applied across different fields such as game theory,
theory, and multi-agent systems. Reinforcement learning is information
methods or algorithms:
further divided into two types of

Positive reinforcement learning: This refers to adding a reinforcing stimulus after a

specificbehavior of the agent, which makes it more likely that the behavior may occur
again in the future, e.g., adding a reward after a behavior.
Negative reinforcement learning: Negative reinforcement learning refers to
strengthening a specific behavior that avoids a negative outcome.