Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Deep Learning Talk - Saverin
Search
Yasser Souri
May 09, 2016
Technology
0
54
Deep Learning Talk - Saverin
Deep Learning Introduction Talk @ Saverin
Yasser Souri
May 09, 2016
Tweet
Share
More Decks by Yasser Souri
See All by Yasser Souri
Intro to Variational AutoEncoder
yassersouri
0
63
Deep Relative Attribute
yassersouri
1
56
Fine-grained Image Classification
yassersouri
1
81
Image Classification Intro
yassersouri
1
130
Real-time tracking of sports pitch markings
yassersouri
1
47
Ensemble of Exemplar-SVMs for Object Detection and Beyond
yassersouri
0
150
Other Decks in Technology
See All in Technology
エンジニアの育成を支える爆速フィードバック文化
sansantech
PRO
3
660
PL900試験から学ぶ Power Platform 基礎知識講座
kumikeyy
0
110
ハッキングの世界に迫る~攻撃者の思考で考えるセキュリティ~
nomizone
12
4.5k
生成AIの利活用を加速させるための取り組み「prAIrie-dog」/ Shibuya_AI_1
visional_engineering_and_design
1
140
急成長する企業で作った、エンジニアが輝ける制度/ 20250214 Rinto Ikenoue
shift_evolve
2
880
Datadogとともにオブザーバビリティを布教しよう
mego2221
0
130
5分で紹介する生成AIエージェントとAmazon Bedrock Agents / 5-minutes introduction to generative AI agents and Amazon Bedrock Agents
hideakiaoyagi
0
220
事業継続を支える自動テストの考え方
tsuemura
0
300
Building Products in the LLM Era
ymatsuwitter
10
4.4k
室長と気ままに学ぶマイクロソフトのビジネスアプリケーションとビジネスプロセス
ryoheig0405
0
320
技術負債の「予兆検知」と「状況異変」のススメ / Technology Dept
i35_267
1
1k
Nekko Cloud、 これまでとこれから ~学生サークルが作る、 小さなクラウド
logica0419
2
730
Featured
See All Featured
The Straight Up "How To Draw Better" Workshop
denniskardys
232
140k
The Language of Interfaces
destraynor
156
24k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
33
2.8k
Mobile First: as difficult as doing things right
swwweet
223
9.3k
Building Better People: How to give real-time feedback that sticks.
wjessup
366
19k
Site-Speed That Sticks
csswizardry
3
370
Statistics for Hackers
jakevdp
797
220k
Building Adaptive Systems
keathley
40
2.4k
The Illustrated Children's Guide to Kubernetes
chrisshort
48
49k
The MySQL Ecosystem @ GitHub 2015
samlambert
250
12k
How to train your dragon (web standard)
notwaldorf
90
5.8k
Adopting Sorbet at Scale
ufuk
74
9.2k
Transcript
Deep Learning Yasser Souri - Alireza Nourian http://sobhe.ir
Have you ever heard of ... Neural Networks
Have you ever heard of ... Deep Learning
Who is he?
Who is he? Jeff Dean, Google
Jeff Dean Creator of Map Reduce, Big Table, Google Crawler
Jeff Dean Creator of Map Reduce, Big Table, Google Crawler
Google Ads, Google Translator, ...
Jeff Dean Facts Compilers don't warn Jeff Dean. Jeff Dean
warns compilers.
Jeff Dean’s Calculator
Jeff Dean’s Current Role Google Brain
DeepMind In 2014, Google acquired DeepMind (a team of ~50)
for ~$ 500 million. And facebook wanted to buy them also.
What is Machine Learning? Problem 1: Given a sequence of
numbers, sort them
What is Machine Learning? Problem 1: Given a sequence of
Farsi characters, output Pinglish
What is Machine Learning? Problem 3: Give a grayscale 28x28
pixel image, identify what number it is.
What is Machine Learning? Problem 3: Give a grayscale 28x28
pixel image, identify what number it is.
What is Machine Learning? x f(x) y Classic
What is Machine Learning? x f(x) y g(x) y’ h(x)
y” Classic
How to Solve Machine Learning Problems Data = (x, y)
Classic
How to Solve Machine Learning Problems Data = (x, y)
y = f(x) Classic (x, y) f(x)
How to Solve Machine Learning Problems Data = (x, y)
y = f(x) Learn the parameters Classic (x, y) f(x; w)
How to Solve Machine Learning Problems Data = (x, y)
y = f(x) Learn the parameters Can x be the raw pixels? Classic (x, y) f(x; w) Features
How to Solve Machine Learning Problems Data = (x, y)
y = f(x) Learn the parameters Can x be the raw pixels? Classic (x, y) f(x; w) Features O(#features) ~ O(#parameters)
Machine Learning Demo http://playground.tensorflow.org/ Classic
Deep Learning Basics Learn from raw data y = f(g(h(
… (x) ))) Deep
Deep Learning Learn from raw data Number of parameters are
much larger y = f(g(h( … (x) ))) Deep
Deep Learning Learn from raw data Number of parameters are
much larger You need more data to learn y = f(g(h( … (x) ))) Deep
Problems being solved with deep learning Deep
Problems being solved with deep learning Deep
One to one: Image Classification Deep
One to one: Image Classification Deep
Problems being solved with deep learning
One to Many: Image Captioning Describing Images:
Fun With ConvNets Describing Images:
Problems being solved with deep learning
May to One: Generating Images Generating Images:
May to One: Generating Images Generating Images:
Problems being solved with deep learning
Statistical Machine Translation
End-to-End Neural Machine Translation (1) Hirschberg, J. & Manning, C.
D. Advances in natural language processing, Science, 2015, 349, 261-266
None
Learning to Execute
Deep Reinforcement Learning
Demo Videos https://www.youtube.com/watch?v=ePv0Fs9cGgU https://www.youtube.com/watch?v=Q70ulPJW3Gk
Fun With ConvNets Modifying images:
Fun With ConvNets Style transfer:
Fun With ConvNets Style transfer:
Fun With ConvNets Colorization:
Fun With ConvNets Colorization:
Fun With ConvNets Colorization:
Fun With ConvNets Colorization:
Fun With ConvNets Colorization:
Fun With ConvNets Colorization:
Fun With ConvNets Colorization:
Growing Use of Deep Learning at Google Jeff Dean &
Oriol Vinyals, “ Large Scale Distributed Systems for Training Neural Networ”, NIPS 2015.
Deep Learning Tools