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
Fader Networks: Manipulating Images by Sliding ...
Search
sobamchan
June 11, 2018
Science
1
310
Fader Networks: Manipulating Images by Sliding Attributes
Fader Networks: Manipulating Images by Sliding Attributes for RINKO
sobamchan
June 11, 2018
Tweet
Share
More Decks by sobamchan
See All by sobamchan
Knowledge Supports Visual Language Grounding_ A Case Study on Colour Terms
sobamchan
1
930
Fast Abstractive Summarization With Reinforce Selected Sentence Rewriting
sobamchan
0
130
Event Representations for Automated Story Generation with Deep Neural Nets
sobamchan
0
78
Other Decks in Science
See All in Science
02_西村訓弘_プログラムディレクター_人口減少を機にひらく未来社会.pdf
sip3ristex
0
520
論文紹介 音源分離:SCNET SPARSE COMPRESSION NETWORK FOR MUSIC SOURCE SEPARATION
kenmatsu4
0
260
Explanatory material
yuki1986
0
350
モンテカルロDCF法による事業価値の算出(モンテカルロ法とベイズモデリング) / Business Valuation Using Monte Carlo DCF Method (Monte Carlo Simulation and Bayesian Modeling)
ikuma_w
0
200
動的トリートメント・レジームを推定するDynTxRegimeパッケージ
saltcooky12
0
160
KH Coderチュートリアル(スライド版)
koichih
1
43k
学術講演会中央大学学員会府中支部
tagtag
0
280
「美は世界を救う」を心理学で実証したい~クラファンを通じた新しい研究方法
jimpe_hitsuwari
1
140
Valuable Lessons Learned on Kaggle’s ARC AGI LLM Challenge (PyDataGlobal 2024)
ianozsvald
0
400
眼科AIコンテスト2024_特別賞_6位Solution
pon0matsu
0
420
ド文系だった私が、 KaggleのNCAAコンペでソロ金取れるまで
wakamatsu_takumu
2
780
白金鉱業Meetup Vol.16_数理最適化案件のはじめかた・すすめかた
brainpadpr
4
1.8k
Featured
See All Featured
Adopting Sorbet at Scale
ufuk
77
9.5k
GraphQLとの向き合い方2022年版
quramy
49
14k
How STYLIGHT went responsive
nonsquared
100
5.6k
Keith and Marios Guide to Fast Websites
keithpitt
411
22k
Evolution of real-time – Irina Nazarova, EuRuKo, 2024
irinanazarova
8
840
Bash Introduction
62gerente
613
210k
Cheating the UX When There Is Nothing More to Optimize - PixelPioneers
stephaniewalter
282
13k
A better future with KSS
kneath
238
17k
The Illustrated Children's Guide to Kubernetes
chrisshort
48
50k
ReactJS: Keep Simple. Everything can be a component!
pedronauck
667
120k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.8k
CSS Pre-Processors: Stylus, Less & Sass
bermonpainter
357
30k
Transcript
Fader Networks: Manipulating Images by Sliding Attributes NIPS 2017 Լ
ᰜଠ
None
֓ཁ 1. Encoder-Decoder FrameworkͰ 2. ը૾ͷಛͷมߋΛ͢ΔϞσϧΛֶश 3. มߋ߹͍ͷௐͰ͖Δ
࣮ߦྫ
શମߏʢޙͰ·ͨͬͯ͘Δʣ
ࣄલࣝ
Encoder-Decoder https://blog.keras.io/building-autoencoders-in-keras.html
GAN (Generative Adversarial Network) https://deeplearning4j.org/generative-adversarial-network
ຊจͷ༰
Encoder-Decoder E Ͱݩը૾ʢxʣΛEncoderͯ͠D Ͱ෮ݩͨ͠ͷͱ ݩը૾ʢxʣͱͰMSEʹΑΓޡࠩΛܭࢉ
Encoder-Decoder E Ͱݩը૾ʢxʣΛEncodeͯ͠D ͰDecodeͨ͠ͷͱ ݩը૾ʢxʣͱͰMSEʹΑΓޡࠩΛܭࢉ ͜ͷ··ͩͱಛͰ͋Δy͕ແࢹ͞ΕΔͷͰ ੑผͷมߋ͕ޙ͔ΒͰ͖ͳ͍
Discriminator objective DiscriminatorEncoderͰಘΒΕͨը૾ͷlatent feature͔Β ͦͷαϯϓϧͷಛ͕ͳΜͳͷ͔ਪఆ͢Δɻ
Adversarial objective EncoderDiscriminatorͷਪఆਫ਼ΛԼ͛ΔΑ͏ʹֶश͢Δ
Adversarial objective EncoderDiscriminatorͷਪఆਫ਼ΛԼ͛ΔΑ͏ʹֶश͢Δ Encoderը૾ͷyʹΑΒͳ͍ ීวతͳಛྔΛநग़͢Δ͜ͱ͕Ͱ͖ΔΑ͏ʹͳΔ
Learning algorithm
શମߏ
Result 1
Result 2
Result 3
·ͱΊ • Encoder-Decoderͷ୯७ͳϑϨʔϜϫʔΫͰը૾ͷಛ มϞσϧΛֶशͨ͠ɻ • Adversarial objectiveΛಋೖ͢Δ͜ͱͰಛʹґଘ͠ͳ ͍ը૾ͷใΛ֫ಘ͢Δ͜ͱ͕Ͱ͖ͨɻ • Networkࣗମͷߏඇৗʹγϯϓϧ͕ͩैདྷख๏Λ্
ճΔΛग़ͨ͠ɻ