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
320
Firebase-React-App
izuna385
0
230
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
520
Unofficial slides: From Zero to Hero: Human-In-The-Loop Entity Linking in Low Resource Domains (ACL 2020)
izuna385
1
630
Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring
izuna385
0
770
Zero-shot Entity Linking with Dense Entity Retrieval (Unofficial slides) and Entity Linking future directions
izuna385
3
980
Entity representation with relational attention
izuna385
0
74
Zero-Shot Entity Linking by Reading Entity Descriptions
izuna385
0
520
Other Decks in Technology
See All in Technology
1等無人航空機操縦士一発試験 合格までの道のり ドローンミートアップ@大阪 2024/12/18
excdinc
0
190
非機能品質を作り込むための実践アーキテクチャ
knih
6
1.7k
10個のフィルタをAXI4-Streamでつなげてみた
marsee101
0
180
なぜCodeceptJSを選んだか
goataka
0
190
終了の危機にあった15年続くWebサービスを全力で存続させる - phpcon2024
yositosi
28
24k
あの日俺達が夢見たサーバレスアーキテクチャ/the-serverless-architecture-we-dreamed-of
tomoki10
0
520
ハイテク休憩
sat
PRO
2
190
生成AIのガバナンスの全体像と現実解
fnifni
1
230
DevFest 2024 Incheon / Songdo - Compose UI 조합 심화
wisemuji
0
200
Server-Side Engineer of LINE Sukimani
lycorp_recruit_jp
0
420
Wantedly での Datadog 活用事例
bgpat
2
820
ISUCON、今年も参加してみた / ISUCON, I challenged it again this year.
dero1to
0
110
Featured
See All Featured
Keith and Marios Guide to Fast Websites
keithpitt
410
22k
Improving Core Web Vitals using Speculation Rules API
sergeychernyshev
1
120
Building Better People: How to give real-time feedback that sticks.
wjessup
366
19k
For a Future-Friendly Web
brad_frost
175
9.5k
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
28
2.1k
Docker and Python
trallard
42
3.2k
[RailsConf 2023] Rails as a piece of cake
palkan
53
5.1k
Fireside Chat
paigeccino
34
3.1k
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
251
21k
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
159
15k
Java REST API Framework Comparison - PWX 2021
mraible
28
8.3k
Embracing the Ebb and Flow
colly
84
4.5k
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