Scribd
Upload
20 views39 pages

PowerPoint Presentation-1

Uploaded by

mubashishola
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
20 views39 pages

PowerPoint Presentation-1

Uploaded by

mubashishola
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/ 39

Fundamentals of

Deep Learning
Part 1: An Introduction to Deep Learning
Agenda
• Part 1: An Introduction to Deep Learning

• Part 2: How a Neural Network Trains

• Part 3: Convolutional Neural Networks

• Part 4: Data Augmentation and Deployment

• Part 5: Pre-Trained Models

• Part 6: Advanced Architectures


To see lecture notes, make full screen
and click the “notes” button
Welcome!
The Goals of This Course

Get you up and on your feet quickly

Build a foundation to tackle a deep learning project right away

We won’t cover the whole field, but we’ll get a great head start

Foundation from which to read articles, follow tutorials, take further classes
Have Fun!
Human Vs Machine Learning
Relaxed Alertness

Human Machine

Rest and Digest Training

Fight-or-flight Prediction
Let’s Get Started
History of AI
Beginning of Artificial Intelligence

Computers are made in Early on, generalized


Turned out to be
part to complete intelligence looked
harder than expected
human tasks possible
Early Neural Networks

• Inspired by biology
• Created in the 1950’s
• Outclassed by Von Neumann Architecture
Expert Systems

Highly complex

Programmed by hundreds of engineers

Rigorous programming of many rules


Expert Systems - Limitations
What are these three images?
How Do Children Learn?

•Expose them to lots of data


•Give them the “correct answer”
•They will pick up the important
patterns on their own
The Deep Learning Revolution
Data

• Networks need a lot of information to learn from


• The digital era and the internet has supplied that data
Computing Power
Need a way for our artificial “brain” to observe lots of data within a practical amount of time.
The Importance of the GPU

A Rendered Image A Neural Network


What is Deep Learning?
Deep learning flips traditional
programming on its head
Traditional Programming
Building a Classifier

Define a set of rules for classification

Program those rules into the computer

Feed it examples, and the program uses


the rules to classify
Machine Learning
Building a Classifier

Show model the examples with the answer


of how to classify

Model takes guesses, we tell it if it’s right


or not

Model learns to correctly categorize as it’s


training. The system learns the rules on its
own
This is a Fundamental Shift
When to Choose Deep Learning

Classic Programming Deep Learning

If rules are clear and straightforward, If rules are nuanced, complex, difficult
often better to program it to discern, use deep learning
Deep Learning Compared to Other AI

Depth and complexity of networks

Up to billions of parameters (and growing)

Many layers in a model

Important for learning complex rules


How Deep Learning is Transforming
the World
Computer Vision

ROBOTICS AND
OBJECT DETECTION SELF-DRIVING CARS
MANUFACTURING
Natural Language Processing

REAL-TIME VOICE
VIRTUAL ASSISTANTS
TRANSLATION RECOGNITION
Recommender Systems

TARGETED
CONTENT CURATION SHOPPING
ADVERTISING
Reinforcement Learning

ALPHAGO BEATS AI BOTS BEAT


STOCK TRADING
WORLD CHAMPION PROFESSIONAL
ROBOTS
IN GO VIDEOGAMERS
Overview of the Course
Hands on Exercises

• Get comfortable with the process of deep learning


• Exposure to different models and datatypes
• Get a jump-start to tackle your own projects
Structure of the Course

“Hello World” of Deep Learning

Train a more complicated model

New architectures and techniques to improve performance

Pre-trained models

Transfer learning
Platform of the Course

GPU powered cloud server

JupyterLab platform

Jupyter notebooks for interactive coding


Software of This Course

• Major deep learning platforms:


• TensorFlow + Keras (Google)
• PyTorch (Meta)
• MXNet (Apache)
• We’ll be using PyTorch
• Good idea to gain exposure to others
moving forward
First Exercise:
Classify Handwritten Digits
Hello Neural Networks

Train a network to correctly classify handwritten digits

Historically important and difficult task for computers

Try learning like a Neural Network

Get exposed to the example, and try to figure out the rules to how it works
Let’s Go!

You might also like