Scribd
Upload
23 views6 pages

Deep Learning: Seungsang Oh

The document provides an introduction to deep learning, including key concepts such as deep neural networks, convolutional neural networks, recurrent neural networks, attention mechanisms, auto-encoders, variational auto-encoders, and generative adversarial networks. It discusses machine learning, artificial intelligence, supervised learning, unsupervised learning, and reinforcement learning. Finally, it outlines major components and milestones in the development of deep learning.

Uploaded by

KaAI Kookmin
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
23 views6 pages

Deep Learning: Seungsang Oh

The document provides an introduction to deep learning, including key concepts such as deep neural networks, convolutional neural networks, recurrent neural networks, attention mechanisms, auto-encoders, variational auto-encoders, and generative adversarial networks. It discusses machine learning, artificial intelligence, supervised learning, unsupervised learning, and reinforcement learning. Finally, it outlines major components and milestones in the development of deep learning.

Uploaded by

KaAI Kookmin
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/ 6

Introduction to Deep Learning

Deep Deep Neural Network


Convolutional Neural Network

Learning Recurrent Neural Network


Attention Mechanism
Seungsang Oh Auto-Encoder & VAE
Dept of Mathematics Generative Adversarial Network
& Dept of Data Science Natural Language Processing
Korea University Graph Neural Network

1 / 279
Artificial Intelligence
• Artificial Intelligence (AI) is the branch of computer science building smart machines
performing tasks that require human intelligence, birthed at Dartmouth Conferences in 1956.
• Machine Learning (ML) is an approach to achieve AI. Rather than hand-coding software,
machines are trained using large data and algorithms and learn how to perform the task.
• Deep Learning (DL) is a technique for implementing ML, thanks to Hinton’s DBN in 2006.
They learn based on huge (deep) neural networks and then run massive amounts of data.

2 / 279
Machine Learning algorithm categories
Supervised Learning
• It learns from a training data set with labels.
• It learns a general rule for regression or classification
to predict the labels for the remaining data.

Unsupervised Learning
• It learns from a training data set without any label.
• It detects patterns in the data by clustering similar
data that have common characteristics.

Reinforcement Learning
• It learns by interacting with the environment, rather than from a sample data set.
• It chooses an action at each state, and receives a reward indicating how good an action is.
Therefore, during training, it adapts its policy in order to maximize the total reward.
• It is often used in robotics, game playing or autonomous cars.

3 / 279
Machine Learning vs Deep Learning
• In Machine Learning, humans extract patterns and co-relation between ‘features’,
but it is difficult to know all the features that would need to be extracted.
• Deep Learning uses many layers to progressively extract higher level features from raw input,
and huge amount of learnable parameters become more accurate as we feed more data,
but many hidden layers lead to a complex structure that is considered as a black box.

4 / 279
Major components in Deep Learning breakthroughs

Big Data Computing Power Algorithms

Deep Neural Network


Convolutional Neural Network
Recurrent Neural Network
Generative Adversarial Network
Deep Reinforcement Learning
Policy Gradient DRL
Bayesian Deep Learning
CPU GPU TPU ..
.

5 / 279
Milestones in Deep Learning developments

6 / 279

You might also like