Scribd
Upload
26 views

Introduction To Neural Networks and Deep Learning

The document provides an introduction to neural networks and deep learning, including definitions of key concepts and how they relate. It explains that neural networks form the foundation of deep learning and machine learning. Deep learning uses complex neural networks with many layers to process large amounts of data without supervision and provide accurate outputs.

Uploaded by

niveditha
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
26 views

Introduction To Neural Networks and Deep Learning

The document provides an introduction to neural networks and deep learning, including definitions of key concepts and how they relate. It explains that neural networks form the foundation of deep learning and machine learning. Deep learning uses complex neural networks with many layers to process large amounts of data without supervision and provide accurate outputs.

Uploaded by

niveditha
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

Introduction to Neural Networks and Deep Learning

Neural Networks are software structures/ algorithms which are designed to imitate the functions of the
human brain.
They have layers of neurons which input data, process them, pass it on to the next hidden layer through
channels and deliver the desired output. This process can be repeated both as a means of forwarding
and reversing the data flow within these layers, until the desired result is achieved.
Neural networks can range from being simple, to medium and even complex structures. They are widely
used in the field of Predictive Modeling. Over the recent decade, the explosion of data into Big Data has
brought with it huge possibilities in the field of Deep Learning. We can store, process, analyze and feed
data and extract information over a wide range of sources. There has been a paradigm shift in the very
structure of data which is pushing the limits of Deep Learning. Take the example of a weather forecast
report. It inputs structured data such as humidity, location, weather history etc.to predict the temperature.
This type of a neural network may need supervision from a domain expert. On the other hand, Youtube
and Amazon recommendations are the sophisticated outputs of complex deep neural networks which
uses specific algorithms to suggest these videos/ products to its users. In this case the Deep Neural
Networks have adapted to understanding the relations between the various labeled and unlabeled data
parameters. Such DNNs may not need supervision as time progresses. Understanding the basic
concepts if Neural networks and Deep learning is one of the premier steps if you are looking for a career
in Artificial Intelligence and Data Science. This free course has been carefully crafted to give a better
insight on the differences between Machine Learning and Deep Learning, Basic set of neural layers, what
is the Activation Function and how it is used to train and achieve accurate output in a neural network. In
addition, users will also be exposed to Deep Learning Algorithm along with a demonstration on the same.
Users will also learn about Convolutional Neural Networks CNNs to understand the application of deep
learning in image processing/ analyzing visual images using deep learning algorithms. This tutorial helps
you delve more into the world of deep learning with respect to understanding industry jargon such as
Recurrent Neural Networks RNNs and Long short-term memory LSTM architecture. By enrolling into this
course, you can as well work on hands-on projects. Upon course completion, a course completion
certificate will be provided which can be showcased across various digital platforms to begin an enriching
and rewarding career in the field of AI.
Subtopics to be covered(FAQs):
- What is neural networks and deep learning? 
Neural networks form the foundation of Deep learning which in turn forms the base of Machine learning
and Artificial Intelligence. Neural Networks are inspired by the structure and functions of the human brain.
They are designed and trained over a period of time to understand patterns, remember their occurrences
by receiving inputs and to eventually produce appropriate outputs. One of the most common application
of Neural Networks is the Facial recognition feature in a mobile phone. Deep learning refers to the use of
algorithms to process information across a set of huge different neural layers within a Neural network.
The data that is processed in Deep Neural Networks can be unstructured and unsupervised as well. 

- Is deep learning and neural networks the same?


No. Deep learning and Neural Networks belong to the world of Artificial Intelligence. They are two
different sub fields of AI. Deep learning refers to how a set of algorithms built on complex neural
networks, processes the input data across neural layers and provides the appropriate output. Neural
network on the other hand refers to the visual representative structure of how these layers are connected
with each other. In other words, Neural Networks form the building blocks of Deep Learning.

- How neural network is used in deep learning? 


Neural networks are the building blocks of Deep learning. Data that is fed to each node in a neural layer is
processed and passed on to the next layer on the neural network. The corresponding output is then
analyzed for its accuracy. If the output is not satisfactory, it could be fed back to the neural network. This
process of ‘training’ the neural network takes place until the desired output is achieved. Neural networks
could be ‘trained’ over a period of time to recognize patterns accurately. In this way, when more neural
layers are added to process unlabeled data, it is said to be a Deep Neural Network or DNN.

- Which are deep neural networks? 


Neural networks usually are built of multiple neural layers. Each of these layers constitute nodes which
process the input data and feeds the information to the next neural layer until it reaches the output layer. 
Deep neural networks process more complex and unstructured data. This makes the number of layers
between the input and output layers huge. Deep neural networks are usually built to process complex
data without any supervision. They act based on previously learned or memorized incidents.

Why use deep neural networks?


Deep neural networks thrive on huge amounts of data that can be processed in its multiple complex
neural layers. In this way, deep neural networks learn incrementally in processing data that is
unstructured or data that needs no supervision. This eliminates the need for frequent intervention during
data processing between the varied neural layers. One can as well achieve near human output accuracy
while using deep neural networks. An example of implementation of deep neural network is that of the
facial recognition feature in a smart phone.
Why do we need deep learning?
It can process huge quantities of data, both structured and unstructured and provide near accurate
results. This feature is used widely in the fields of predictive modeling, Medicine, Technology, Robotics,
Entertainment etc. Example, YouTube’ personalized video recommendations, which is one of the largest
in the entertainment industry. Thanks to deep learning, now YouTube can better understand and learn to
suggest videos for every user.
What exactly is deep learning?
Deep learning uses deep neural networks DNNs that are complex and have huge neural layers. These
neural layers learn to predict more accurate results based on the large amounts of unstructured data that
is fed into it. This type of learning by the neural network which is more complex in structure and which
can deliver the output without any kind of intervention, as time progresses is called deep learning. In other
words, the system learns on its own to solve problems and deliver the desired output over a period of
time. One of the widely used fields of deep learning is Robotics.
How do deep neural networks work?
DNNs are basically built of 3 types of neural layers to process data and provide the desired result. These
are the Input layer, the Hidden layers and the Output layer.Each layer consists of a number of
interconnected data neurons. The data from the input layer is processed and passed on to the hidden
layer where more data processing takes place. The weight value associated with neurons help in the
derivation of Activation Function which in turn helps achieve near-perfect output accuracy. This way, the
neural network gets trained progressively, eventually to become self-taught about that domain and work
independently without any supervision. Example, teaching and training a Humanoid to identify and
remember colors, numbers, alphabets etc.During the initial stages it would need supervision but as time
progresses it would learn to do so on its own.

You might also like