Deep learning vs.

Machine
learning vs. Artificial Intelligence

• Deep Learning, Machine Learning, and Artificial Intelligence are the most used
terms on the internet for IT folks. However, all these three technologies are
connected with each other.
• Artificial Intelligence (AI) can be understood as an umbrella that consists of both
Machine learning and deep learning.
What is Artificial Intelligence (AI)?
• Artificial Intelligence is defined as a field of science and engineering
that deals with making intelligent machines or computers to perform
human-like activities.
• Mr. John McCarthy is known as the godfather of this amazing
invention.
What is Machine Learning?
• Machine Learning is defined as the branch of Artificial Intelligence and
computer science that focuses on learning and improving the
performance of computers/machines through past experience by
using algorithms.
• AI is used to make intelligent machines/robots, whereas machine
learning helps those machines to train for predicting the outcome
without human intervention.
How does Machine Learning work?
• Machine Learning uses algorithms and techniques that enable the
machines to learn from past experience/trends and predict the
output based on that data.
Types of Machine Learning
Based on the methods and techniques to teach machines, Machine Learning is categorized into mainly four
types, which are as follows:
Supervised Machine Learning
• This type of ML method uses labeled datasets to train machines and, based on
these datasets, machines predict the output. It needs supervision to train models
and predict outputs.
• Image segmentation, medical diagnosis, fraud detection, spam detection,
speech recognition, etc., are some important applications of supervised machine
learning.

• Supervised machine learning can be further categorized into 2 types of problems as

follows:
Classification
Regression
Advantages of Supervised machine learning
•Supervised machine learning helps to predict output based on
prior experience
•It helps to provide an exact idea about classes of objects.
Disadvantages of Supervised machine learning
•This method is not significant in solving complex problems.
•This method does not guarantee to give exact output as it
contains both structured and unstructured data.
•It needs more computational time to teach ML models.
Unsupervised Machine Learning
• Unsupervised machine learning is just the opposite of supervised learning. Unlike
supervised machine learning, it does not need supervision, which means it does not
require labeled datasets to train machines

• Hence, in unsupervised machine learning, the output is predicted without any

supervision. The main aim of the unsupervised learning algorithm is to group or
categorize the unsorted dataset according to the similarities, patterns, and differences.
• Network analysis, recommendation system, anomaly detection, singular value
decompositions, etc., are some important applications of unsupervised machine
learning.

• Unsupervised machine learning is further categorized into two types:

Clustering
Association
Advantages of unsupervised machine learning
• It can be used to solve complex ML problems as it works with
unlabelled data sets.
• It is used to solve multiple tasks in comparison to supervised learning.
Disadvantages of unsupervised machine learning
• Using unlabeled data sets may predict inaccurate outputs.
• It is a relatively complex algorithm as it deals with unlabelled datasets
and also does not map with output.
Semi-supervised Machine learning
Semi-supervised learning is the combination of both supervised and
unsupervised machine learning.
•Although it uses both labeled and unlabelled datasets to train models and predict
the output, mostly, it contains the unlabelled datasets.

Advantages of Semi-supervised machine learning

•It is simple and easy to understand the algorithm.
•It is more efficient.
•It is used to solve the drawbacks of Supervised and Unsupervised Learning
algorithms.

Disadvantages of Semi-supervised machine learning

•It does not include applicable network-level data
•It gives less accurate results
•Iterations results may not be stable.
Reinforcement Learning
• Reinforcement learning is defined as the feedback-based method to learn from past experience and
improve the performance of models. In this method, an AI agent automatically explores its surrounding by
hitting and trial actions.
• Further, in reinforcement learning algorithms, machines learn from experience or past data and do not use
labeled data. It can be applied in various real-world cases such as video games, resource management,
robotics, text mining, operations & research, etc.

Advantages of reinforcement learning

•It is used to resolve complex real-time scenarios where all other techniques are not useful.
•It provides the most accurate results because it learns similarly to a human.
•It is significant for achieving long-term results.

Disadvantages of Reinforcement Learning

•It is not significant for simple scenarios.
•It needs a vast amount of data as well as computations.
Steps involved in machine learning
• There are 7 simple steps involved in machine learning as follows:
• Data gathering
• Data pre-processing
• Choose model
• Train model
• Test model
• Tune model
• Prediction
What is Deep Learning?

• "Deep learning is defined as the subset of machine learning and artificial intelligence that is based on
artificial neural networks". In deep learning, the deep word refers to the number of layers in a neural
network.
• Deep Learning is a set of algorithms inspired by the structure and function of the human brain. It uses
a huge amount of structured as well as unstructured data to teach computers and predicts accurate
results. The main difference between machine learning and deep learning technologies is of
presentation of data. Machine learning uses structured/unstructured data for learning, while deep
learning uses neural networks for learning models.
• In machine learning, if a model predicts inaccurate results, then we need to fix it manually. Further, in
deep learning techniques, these problems get fixed automatically, and we do not need to do anything
explicitly. A self-driving vehicle is one of the best examples to understand deep learning.
• Deep learning can be useful to solve many complex problems with more accurate predictions such as
image recognition, voice recognition, product recommendations systems, natural language
processing (NLP), etc.
The basic structure of deep learning
• Deep learning includes various neural networks that possess different
layers, such as input layers, hidden layers, and output layers. The input
layer accepts input data; hidden layers are used to find any hidden pattern
and feature from the data, and output layers show the expected results.
How does deep learning work?
There are a few simple steps that deep learning follows.
1.Calculate the weighted sum
2.Use this weighted sum in step1 as input for the activation function.
3.The activation function adds bias and decides whether the neuron
should be triggered or not.
4.Predict output at the output layer.
5.Compare predicted output and actual output and accordingly use the
backpropagation method for improving the performance of the
model. In this step, the cost function plays a vital role in reducing the
error rate.
Types of deep neural networks
There are some different types of deep learning networks available. These
are as follows:
• Feedforward neural network
• Radial basis function neural networks
• Multi-layer perceptron
• Convolution neural network (CNN)
• Recurrent neural network
• Modular neural network
• Sequence to sequence models
Applications of Deep Learning:
• Self-driving vehicles
• Fraud detection
• Natural language processing
• Virtual personal assistance
• Text, speech, and image recognition
• Healthcare, infrastructure, banking & finance, marketing
• Entertainment
• Education
• Automatic game playing
• Auto handwriting generation
• Automatic language translation
• Pixel restoration and photo description & tagging
• Demographic and election predictions, etc.
S. No. Machine Learning Deep Learning

Machine Learning is a superset of Deep Learning is a subset of Machine

1.
Deep Learning Learning

The data represented in Machine

The data representation is used in
Learning is quite different as compared
2. Deep Learning is quite different as it
to Deep Learning as it uses structured uses neural networks(ANN).
data

Deep Learning is an evolution of

3. Machine Learning is an evolution of AI Machine Learning. Basically, it is how
deep is the machine learning.

Machine learning consists of

4. Big Data: Millions of data points.
thousands of data points.

Anything from numerical values to

5. Outputs: Numerical Value, like free-form elements, such as free text
classification of the score.
and sound.
 Uses various types of automated Uses neural network that passes data
6. algorithms that turn to model functions through processing layers to, interpret
and predict future action from data. data features and relations.

Algorithms are detected by data analysts Algorithms are largely self-depicted on

7. to examine specific variables in data data analysis once they’re put into
sets. production.

Machine Learning is highly used to stay

Deep Learning solves complex machine
8. in the competition and learn new learning issues.
things. 

Training can be performed using the A dedicated GPU (Graphics Processing

9.
CPU (Central Processing Unit). Unit) is required for training.

Although more difficult to set up, deep

10. More human intervention is involved for learning requires less intervention once
getting results.
it is running.
 Although they require additional setup
Machine learning systems can be swiftly time, deep learning algorithms can
11. set up and run, but their effectiveness produce results immediately (although
may be constrained.  the quality is likely to improve over time
as more data becomes available).

Its model takes less time in training due A huge amount of time is taken because
12.
to its small size. of very big data points.

Feature engineering is not needed

Humans explicitly do feature because important features are
13.
engineering. automatically detected by neural
networks.

Machine learning applications are Deep learning systems utilize much

14. simpler compared to deep learning and
can be executed on standard computers. more powerful hardware and resources.