Scribd
Upload
23 views20 pages

Deep Learning with R Overview

The document provides an overview of a deep learning course that focuses on both theory and applications using Python/R and frameworks like PyTorch and Keras. It discusses the history of artificial intelligence, the distinction between machine learning and deep learning, and highlights various applications of deep learning in areas such as computer vision, natural language processing, and medical imaging. Additionally, it outlines the reasons for the recent advancements in deep learning, including improvements in hardware, availability of large datasets, and algorithmic developments.

Uploaded by

okuwobi
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
23 views20 pages

Deep Learning with R Overview

The document provides an overview of a deep learning course that focuses on both theory and applications using Python/R and frameworks like PyTorch and Keras. It discusses the history of artificial intelligence, the distinction between machine learning and deep learning, and highlights various applications of deep learning in areas such as computer vision, natural language processing, and medical imaging. Additionally, it outlines the reasons for the recent advancements in deep learning, including improvements in hardware, availability of large datasets, and algorithmic developments.

Uploaded by

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

Deep Learning

Introduction
Adapted from DLP Book Chapter 1.
Course Overview
• Focus on both theory and applications

• Programming language: python/R + pytorch (keras)

• References:
• Ian Goodfellow’s book on deep learning
• Deep learning with Python by Francois Chollet
• Deep learning with R by Francois Chollet
• This book is kind of a translation from the 2nd book into R.
What is deep learning
• Deep learning a brief history
• Artificial intelligence:
• Emerged in the 1950s; focused on the question of how to make computers
think
• Goal is to automate intellectual tasks commonly done by humans
• It used to be believed that AI could be achieved using a set of detailed enough
rules
• This is called symbolic AI.
Machine Learning
• A precursor for “machine” is what is called “analytical machine”
• By Ada Lovelace and Charles Babbage (mid Victoria era).
• A general purpose machine used to automate repetitive steps;

• Turing machine is the beginning of modern computers.


• In addition to being told the rules to repeat, machines could learn new rules.
• This is what distinguishes programming language from a machine learning
algorithm
Deep Learning and Machine
Learning
• Machine learning tries to learn useful ways to “represent” the input data so
that it could be “transformed” to output that we want.
• A toy example of ‘representing the data’.

• Deep learning seeks to learn multiple layers of representation


• “deep” means many layers.
Deep Learning Paradigm
• The data transformation is stored in the
weights
• How well the transformation works is
measured by the loss function
• Weights can be updated to make the
transformation better
Deep Learning’s Achievement So Far
• There are many deep learning applications that are in our daily life now:
• Things that DL do almost equally well as humans
• Computer vision
• Face recognition
• Speech recognition/dictation
• Translation
• Autonomous driving
• Things that DL do better than human
• Playing GO
• Pattern recognition
• Medical image processing
Computer Vision
• Image classification
• Cat or Dog

More
challenging
Computer Vision
• Object detection (putting a box)
• Image segmentation (including boundary detection as well)
Natural language processing
• Translation

• Language models

• Captioning
Medical applications
• AliveCor from EKG T wave to potassium levels
Apple watch and AF
Medical Imaging
Some other recent deep learning
news
Why Deep Learning Works Now
• Most of the algorithms exist since 1970s, and some in late 1990s.
• Many of the algorithm remain unchanged ever since
• E.g. LSTM

• Three major advances make DL a cutting edge development


• Hardware
• GPU
CPU vs. GPU
GPU machine in my lab
Why Deep Learning Works Better
Now
• Large collection of labelled data
• ImageNet
• Kaggle data mining challenge

• Algorithms
• Activation function (ReLU)
• Algorithms that help convergence
• Adam
• RMSProp
• $$
Topics to be covered
• Deep forward network

• Convoluted neural networks


• Mostly developed for imaging analysis
• Sequence models
• RNN
• Attention and transformer.
• Deep generative models
• GANs

You might also like