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
Problems of Neural Networks and its solutions
Search
izuna385
June 21, 2018
Technology
0
120
Problems of Neural Networks and its solutions
Residual Connections とBatch Normalizationがメイン
izuna385
June 21, 2018
Tweet
Share
More Decks by izuna385
See All by izuna385
jel: japanese entity linker
izuna385
0
300
Firebase-React-App
izuna385
0
220
React+FastAPIを用いた簡単なWebアプリ作製
izuna385
0
1.5k
UseCase of Entity Linking
izuna385
0
470
Unofficial slides: From Zero to Hero: Human-In-The-Loop Entity Linking in Low Resource Domains (ACL 2020)
izuna385
1
610
Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring
izuna385
0
710
Zero-shot Entity Linking with Dense Entity Retrieval (Unofficial slides) and Entity Linking future directions
izuna385
3
930
Entity representation with relational attention
izuna385
0
73
Zero-Shot Entity Linking by Reading Entity Descriptions
izuna385
0
480
Other Decks in Technology
See All in Technology
内製化を目指す事業会社が、システム開発会社と共に進める「開発生産性改善」の取り組み事例 #devsumi
yuwji
0
100
「家族アルバム みてね」における運用管理・ オブザーバビリティの全貌 / Overview of Operation Management and Observability in FamilyAlbum
isaoshimizu
4
160
OSTという文化を組織に根付かせてみた
sansantech
PRO
2
390
とあるOSSを継続可能にするための取り組みについて / OSS Refactoring Process
bun913
1
210
やってやろうじゃないかメカアジャイル! / Let's do it, mechanical agile!
psj59129
1
670
ナレッジグラフとLLMの相互利用
koujikozaki
0
420
Mocking in Rust Applications
taiki45
2
410
Road to Single Activity
yurihondo
2
240
可視化により内部品質をあげるAIドキュメントリバース/20240910 Hiromitsu Akiba
shift_evolve
0
220
2024年のナビゲーション・フォーカス対応:Composeでキーボード・ナビゲーションをサポートしよう
tahia910
0
110
watsonx.ai Dojo 環境準備について
oniak3ibm
PRO
0
300
Discovering AI Models
picardparis
4
3.9k
Featured
See All Featured
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.2k
What's in a price? How to price your products and services
michaelherold
242
11k
KATA
mclloyd
27
13k
A Tale of Four Properties
chriscoyier
155
22k
How to Think Like a Performance Engineer
csswizardry
16
960
[RailsConf 2023 Opening Keynote] The Magic of Rails
eileencodes
28
8.9k
Put a Button on it: Removing Barriers to Going Fast.
kastner
58
3.4k
Ruby is Unlike a Banana
tanoku
96
11k
The Language of Interfaces
destraynor
153
23k
Atom: Resistance is Futile
akmur
261
25k
The Mythical Team-Month
searls
218
43k
Web development in the modern age
philhawksworth
205
10k
Transcript
1 / 18 Neural Networks
2 / 18 1. NN !
• Residual Network • Batch Normalization 2. 1. • •
3 / 18 Plain NNs(&) ' pros #%
" (ex. CNN, RNN, ...) cons ! $ $
4 / 18 RNN RNN [1] P. Razvan et
al ,"On the difficulty of training recurrent neural networks." International Conference on Machine Learning. 2013. !"#$ !" %"&$ %"#$ %" %"&$ '() '() '() '*+, '*+, -!"# = /(!!"# ) -! -!$# %! : input !! : hidden state '%&' : '() : input / !" = '*+, 2 !"#$ + '() %"
5 / 18 !" !# !$ %" %# %$ &'(
&'( &'( &)*+ &)*+ ,! = .(!! ) ," ,# RNN 3 1, 12 = 1," 12 + 1,# 12 + 1,$ 12 1,$ 12 = 4 "565$ 1,$ 1!$ 7 1!$ 1!6 7 18!6 12 1!$ 1!" = 1!$ 1!# 7 1!# 1!" = &)*+ 9 :;<= >? !# 7 &)*+ 9 :;<= >? !" @A!B @C : !" ~!6E" fix !6
6 / 18 RNN Vanishing/Exploding Gradient : !"#$ !%&
'( )( … … )* '* ………… ………… +( +* !"#$ (-) !%& (-) '% …… '/ )/ +/
7 / 18 ,$+ /' !"#$ !- !"#$ 2 %
× '()* + ×%,- → # !"#$ !"#$ . 2 % × '()*(+).,-×%,- 1%input or 1)* Loss( RNN ."0& Vanishing/Exploding Gradient
8 / 18 +$ DeepNN( ! +
" )*&!/#% ' (→ ! Loss func ! Loss func → Residual Connection, Batch No malization
9 / 18 0), : Residual Connection – -– F(x)
"/#2 → "/ F(x) + x → (4 '$"/ Identity Mapping +%*1&: 3 . ! 3 Identity – [1] He, Kaiming, et al. "Identity mappings in deep residual networks." European Conference on Computer Vision. Springer, Cham, 2016.
10 / 18 : Residual Connection –– ' Forward
$#& Backward !$"& Deep % & input
11 / 18 Residual Connection –– https://icml.cc/2016/tutorials/icml2016_tutorial_deep_residual_networks_kaiminghe.pdf
12 / 18 ResNet Batch Normalization ResNet Residual Block
• ImplementationBatch Normalization NN ! $# • Batch Normalization" ## http://torch.ch/blog/2016/02/04/resnets.html Plain
13 / 18 ( ) 1 2
( ) n … Batch Normalization –Revisit Gaussian-
14 / 18 Batch Normalization -Input Data distribution
- (Convergence) !! Input NN → input
15 / 18 Batch Normalization -distribution - !"#$% & '
= ) & ' ← ' − , - ~/(,, -2) input
16 / 18 Batch Normalization Data distribution •
=(!, ")fix • Batch Normalization Batch Normalization
17 / 18 Batch Normalization – [2]Ioffe, Sergey,
and Christian Szegedy. "Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift." (2015). !, # !%$( → normalize scaling '"&# nomalize
18 / 18 DeepNN+ ! /
& -"#.#)%/'( *$ +!→ , Identity – normalize scaling implement Deep Net