Scribd
Upload
7 views19 pages

DNN Intro

Deep learning is a subset of machine learning that uses multi-layered neural networks to analyze both structured and unstructured data, enabling systems to make predictions with high accuracy. It has applications in various fields such as virtual assistants, healthcare, and image processing, outperforming traditional machine learning in handling large datasets and complex tasks. Key differences between machine learning and deep learning include data dependency, execution time, and the need for feature engineering.

Uploaded by

Karan Sharma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
7 views19 pages

DNN Intro

Deep learning is a subset of machine learning that uses multi-layered neural networks to analyze both structured and unstructured data, enabling systems to make predictions with high accuracy. It has applications in various fields such as virtual assistants, healthcare, and image processing, outperforming traditional machine learning in handling large datasets and complex tasks. Key differences between machine learning and deep learning include data dependency, execution time, and the need for feature engineering.

Uploaded by

Karan Sharma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

Deep Learning & AI

Dr Akhter Mohiuddin
Deep learning

Deep learning attempts to Deep learning is a subset of machine


mimic the human brain— learning, which is essentially a neural
albeit far from matching its network with three or more layers. These
ability—enabling systems to neural networks attempt to simulate the
cluster data and make behavior of the human brain—albeit far
predictions with incredible from matching its ability—allowing it to
accuracy. “learn” from large amounts of data.

2
AI vs ML vs DL
Deep learning & Neural networks

⬢ Deep learning algorithms attempt


to draw similar conclusions as
humans.
⬢ To achieve this, DL uses a multi-
layered structure of algorithms
called neural networks.

4
Importance of DL
⬢ ML works only with sets of structured and semi-structured data, while deep learning
works with both structured and unstructured data
⬢ Deep learning algorithms can perform complex operations efficiently, while machin
learning algorithms cannot
⬢ Machine learning algorithms use labeled sample data to extract patterns, while dee
learning accepts large volumes of data as input and analyzes the input data to extr
features out of an object
⬢ The performance of machine learning algorithms decreases as the number of data
increases; so to maintain the performance of the model, we need a deep learning

5
Deep
learning
Applications
6
Applications

⬢ Virtual Assistants
⬢ Chatbots
⬢ Healthcare
DL is used for medical
research, drug discovery,
and diagnosis of life- ⬢ Companies such as Netflix, Amazon, YouTube, and Spotify give relevant
threatening diseases. movies, songs, and video recommendations to enhance their customer
experience.
⬢ Based on a person’s browsing history, interest, and behavior, online
streaming companies give suggestions to help them make product and
service choices.

7
Image Coloring Image Captioning

⬢ Image Captioning is the method of


⬢ Image colorization is taking an input generating a textual description of
of a grayscale image and then an image.
producing an output of a colorized ⬢ It uses computer vision to
image. understand the image's content
⬢ ChromaGAN is an example of a and a language model to turn the
picture colorization model. understanding of the image into
words in the right order.
8
Deep
learning
Image
Classification
9
Image classification is
where a computer can
analyse an image and
identify the ‘class’ the
image falls under.

10
Image
Classification
Datasets

11
CIFAR 10 MNIST Handwritten
digits
MNIST Fashion

12
DL and
Natural
Language
Processing
13
Applications in NLP
⬢ Search Autocorrect and ⬢ Social media posts some ⬢ Sentiment analysis for text

Autocomplete useful information about with Deep Learning


an individual’s likes and
dislikes.
⬢ Analyzing this unstructured
data can help in generating
valuable insights.

14
Parameter ML DL
Data Dependency ⬢ ML depends on the huge ⬢ DL depends on a large
amount of data, it can still amount of data, so we
work on small amount of need to feed a large
data. amount of data for good
performance.
Execution time ⬢ ML takes less time to ⬢ DL takes a long
train the model than deep execution time to train
learning, but it takes a the model, but less time
long-time duration to test to test the model.
the model.

15
Parameter ML DL
Hardware ⬢ ML models do not need ⬢ DL model needs a huge
Dependencies much amount of data, so amount of data to work
they can work on low-end efficiently, so they need
machines. GPU's and hence the
high-end machine.
⬢ ML models need a step of ⬢ DL does not need to
Feature
feature extraction by the develop the feature
Engineering
expert, and then it extractor for each
proceeds further. problem; instead, it tries
to learn high-level
features from the data on
its own.
16
Parameter ML DL
Type of data ⬢ ML models mostly require ⬢ DL models can work with
data in a structured form. structured and
unstructured data both
⬢ ML models are suitable as they rely on the layers
Suitable for for solving simple or bit- of the ANN.
complex problems.
⬢ DL models are suitable
for solving complex
problems.

17
Deep Neural Network

18
Contact:

[email protected]

19

You might also like