Upgrade to PRO for Only $50/Year—Limited-Time Offer! 🔥
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
[Journal club] Scalable Diffusion Models with T...
Search
Semantic Machine Intelligence Lab., Keio Univ.
PRO
July 22, 2024
Technology
0
77
[Journal club] Scalable Diffusion Models with Transformers
Semantic Machine Intelligence Lab., Keio Univ.
PRO
July 22, 2024
Tweet
Share
More Decks by Semantic Machine Intelligence Lab., Keio Univ.
See All by Semantic Machine Intelligence Lab., Keio Univ.
Will multimodal language processing change the world?
keio_smilab
PRO
2
260
[Journal club] MOKA: Open-Vocabulary Robotic Manipulation through Mark-Based Visual Prompting
keio_smilab
PRO
0
41
[Journal club] Seeing the Unseen: Visual Common Sense for Semantic Placement
keio_smilab
PRO
0
43
[Journal club] Language-Embedded Gaussian Splats (LEGS): Incrementally Building Room-Scale Representations with a Mobile Robot
keio_smilab
PRO
0
42
[Journal club] RAM: Retrieval-Based Affordance Transfer for Generalizable Zero-Shot Robotic Manipulation
keio_smilab
PRO
1
55
[Journal club] Simplified State Space Layers for Sequence Modeling
keio_smilab
PRO
0
36
[Journal club] Detecting and Preventing Hallucinations in Large Vision Language Models
keio_smilab
PRO
1
84
[IROS24] Object Segmentation from Open-Vocabulary Manipulation Instructions Based on Optimal Transport Polygon Matching with Multimodal Foundation Models
keio_smilab
PRO
0
57
[IROS24] Learning-To-Rank Approach for Identifying Everyday Objects Using a Physical-World Search Engine
keio_smilab
PRO
0
92
Other Decks in Technology
See All in Technology
LLMアプリケーションの評価と継続的改善
pharma_x_tech
3
180
深層学習のリペア技術の最新動向と実際 / DNN Repair Techniques for AI Performance Alignment for Safety Requirements
ishikawafyu
0
140
高品質と高スピードを両立させるソフトウェアQA/Software QA that Supports Agility and Quality
goyoki
4
460
LINEヤフーにおける超大規模プラットフォーム実現への挑戦と学び / Challenges and Lessons in Building an Ultra-Large-Scale Platform at LY Corporation
hhiroshell
2
950
累計2500万着電を支える大規模 電話自動応答サービスのアーキテクチャ / Architecture of a Large-Scale Automated Phone Response Service Supporting 25 Million Cumulative Calls
ymachida
8
4.3k
Bytebaseで実現する データベース管理の効率化
shogo452
1
300
専門領域に特化したチームの挑戦
leveragestech
0
240
エンジニアの草の根活動のその先へ LINEギフトのアクセシビリティにおける ネクストアクション
lycorptech_jp
PRO
0
110
全社員に向けて生成AI活用を促進!~電通総研の生成AI活用ロードマップ~
iotcomjpadmin
0
310
セキュリティベンダー/ユーザー双方の視点で語る、 Attack Surface Managementの実践 - Finatextパート / cloudnative-architecture-of-asm
stajima
0
2.6k
AWS認定試験の長文問題を早く解くコツ
keke1234ke
0
140
LLMを「速く」「安く」 動かすには / CloudNative Days Winter 2024
pfn
PRO
5
1.2k
Featured
See All Featured
Code Reviewing Like a Champion
maltzj
520
39k
RailsConf 2023
tenderlove
29
920
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
27
2.1k
StorybookのUI Testing Handbookを読んだ
zakiyama
27
5.3k
Mobile First: as difficult as doing things right
swwweet
222
8.9k
Visualization
eitanlees
145
15k
Art, The Web, and Tiny UX
lynnandtonic
297
20k
Scaling GitHub
holman
458
140k
Measuring & Analyzing Core Web Vitals
bluesmoon
4
150
Producing Creativity
orderedlist
PRO
341
39k
Side Projects
sachag
452
42k
Building a Modern Day E-commerce SEO Strategy
aleyda
38
7k
Transcript
4DBMBCMF%JGGVTJPO.PEFMTXJUI 5SBOTGPSNFST ܚጯٛक़େֶ ਿӜ໌ݚڀࣨ#ീౡେ 8JMMJBN1FFCMFT 4BJOJOH9JF 6$#FSLFMFZ /FX:PSL6OJWFSTJUZ *$$7 8JMMJBN1FFCMFT
4BJOJOH9JFl4DBMBCMF%JGGVTJPO.PEFMTXJUI5SBOTGPSNFSTzJO*$$7 QQ
എܠɿ֦ࢄϞσϧʹΑΔಈը૾ੜ FH 4PSB ͷൃల IUUQTXXXZPVUVCFDPNXBUDI W),Z%"1/@
ؔ࿈ݚڀɿ֦ࢄϞσϧͷόοΫϘʔϯͱͯ͠6/FU͕ଟ༻ • 6/FUͷ.VMUJTDBMFTLJQDPOOFDUJPOTˠ ෆཁͳܭࢉࢿݯͷ༻ 手法 概要 DALL-E 2 [Ramesh+,
22] CLIPを用いてテキストと画像のAlignmentを行う Stable Diffusion [Rombach+, CVPR22] 潜在拡散モデル 6/FU<3POOFCFSHFS .*$$"*> 4UBCMF%JGGVTJPO<3PNCBDI $713>
ఏҊख๏ɿ%JGGVTJPO5SBOTGPSNFS %J5 • જࡏ֦ࢄϞσϧ -%. <3PNCBDI $713>Λϕʔεʹߏங • 7JTJPO5SBOTGPSNFS 7J5
<%PTPWJUTLJZ *$-3>ػߏΛಋೖ • $POEJUJPOJOHʹΑΔ݅ใͷೖྗ
ఏҊख๏ ɿજࡏ֦ࢄϞσϧͱֶͯ͠शͤ͞Δ • ߴ࣍ݩͷըૉۭؒͰ֦ࢄϞσϧΛֶश ͤ͞Δ͜ͱܭࢉྔతʹࠔ • -%.ͱֶͯ͠शͤ͞Δ͜ͱͰ ܭࢉྔΛݮ
• ըૉۭؒͷ֦ࢄϞσϧͰ͋Δ "%.<%IBSJXBM /FVS*14> ͷͷͷ(GMPQTͰֶशՄೳ
ఏҊख๏ ɿೖྗϊΠζΛQBUDIʹղ • "VUPFODPEFS͔ΒಘΒΕͨ /PJTFE-BUFOU YY Λ 7J5ͱಉ༷ʹE࣍ݩͷτʔΫϯ5ʹม •
1BUDIαΠζQΛʹ͢Δͱ5ഒ ʹͳΓUSBOTGPSNFSͷ (GMPQTগͳ͘ͱഒҎ্
ఏҊख๏ ɿ͖݅ೖྗ $POEJUJPOJOH ͷॲཧ • ͖֦݅ࢄϞσϧͰϊΠζΛؚΉը૾ͱͱʹՃใ͕Ճ͑ΒΕΔ FH UJNFTUFQɼΫϥεϥϕϧɼࣗવݴޠ FUD
• ຊݚڀͰ͜ΕΒͷ͖݅ೖྗΛॲཧ͢ΔͨΊʹҎԼͷͭͷҟͳΔઃܭΛఏҊ • *ODPOUFYUDPOEJUJPOJOH • $SPTT"UUFOUJPOCMPDL • "EBQUJWFMBZFSOPSN BEB-/ CMPDL • BEB-/;FSPCMPDL
ఏҊख๏ ɿBEB-/;FSPCMPDL • 7J5ͷTFMGBUUFOUJPOCMPDLʹରͯ͠"EB-/ػߏΛಋೖ • "EB-/ͷεέʔϧ ͓Αͼ ࠩଓͷલͷεέʔϧ
Λύϥϝʔλͱͯ͠Ճ ˠ݅ใΛը૾ʹΑΓڧ͘ө • "EB-/;FSPCMPDLͰͦΕΒΛθϩʹॳظԽ ˠֶशͷॳظஈ֊߃ؔʹ͍ۙಇ͖ ˠ ֶशͷ҆ఆԽ
࣮ݧઃఆ • σʔληοτ • $MBTT$POEJUJPOBM*NBHF/FUY Y<%FOH $713> • ΞʔΩςΫνϟ
• 7J5ͱಉ༷ʹͭͷϞσϧͷେ͖͞ 4 # - 9- Λ༻ҙ • QBUDITJ[FQ • %%1.TBNQMJOHTUFQT • ධՁई • '*% T'*% *4 1SFDJTJPO 3FDBMM • (GMPQT • ֶश • 516WQPE #BUDITJ[F
ఆྔత݁Ռɿ6/FUϕʔεͷख๏Λ্ճͬͨ
ఆੑత݁Ռ • 1BUDITJ[FΛখ͘͞ɼϞσϧΛେ͖͘͢ΔͱΑΓࣗવͳը૾͕ग़ྗ͞ΕΔ ˠ%J5Ͱ(GMPQT͕େ͖͍΄Ͳग़ྗը૾ͷ্࣭͕͕Δ
ࢼ͓ΑͼΤϥʔੳ ఆੑత݁Ռ ɿࣦഊྫ • ಛఆͷMBCFMʹରͯ͠ෆࣗવͳը૾͕ੜ͞ΕΔ • ྫɿJOQVUMBCFM UPZQPPEMF %%1.TBNQMJOHTUFQ
• ϥϕϧʹΑͬͯTUFQͰੜը૾͕ෆ҆ఆˠ ਪ࣌ͷTUFQΛಈతʹมߋ
ॴײ • 4USFOHUI • ֦ࢄϞσϧʹUSBOTGPSNFSΛಋೖ • ܭࢉࢿݯͱग़ྗը૾ͷ࣭ʹ͍ͭͯͷߟ • 8FBLOFTT
• ͕ࣜগͳ͔ͬͨ • Τϥʔੳ͕ͳ͍
·ͱΊ • എܠ • ֦ࢄϞσϧʹΑΔಈը૾ੜ FH 4PSB ͷൃల • ֦ࢄϞσϧʹ͓͚ΔUSBOTGPSNFSͷར༻͕গͳ͍
• ఏҊख๏ • USBOTGPSNFSϕʔεͷ֦ࢄϞσϧͰ͋Δ%JGGVTJPO5SBOTGPSNFS %J5 ΛఏҊ • ݁Ռ • %J5εέʔϥϏϦςΟ͕ߴ͘ɼ(GMPQT͕େ͖͍΄Ͳ'*%͕Լ ˠ ܭࢉࢿݯͱग़ྗը૾ͷ࣭ʹڧ͍૬ؔؔ • %J59-Ϟσϧ͕ɼ$MBTT$POEJUJPOBM*NBHF/FUʹ͓͍ͯ ैདྷͷ6/FUϕʔεͷ֦ࢄϞσϧΛ্ճͬͨ
"QQFOEJYɿ%FOPJTJOH%JGGVTJPO1SPCBCJMJTUJD.PEFM %%1. ֶश
"QQFOEJYɿ$MBTTJGJFSGSFFHVJEBODF • ͖֦݅ࢄϞσϧͰΫϥεϥϕϧΛϥϯμϜʹυϩοϓ ˠ αϯϓϦϯάͷਫ਼Λ্ • #BZFTͷఆཧΑΓ • ֦ࢄϞσϧͷग़ྗΛείΞͱͯ͠ղऍ͢Δͱਪఆ͢ΔϊΠζҎԼͷΑ͏ʹͳΔ
TɿΨΠμϯεεέʔϧ
"QQFOEJYɿ*ODPOUFYUDPOEJUJPOJOH • $POEJUJPOJOHͰ݅ͱͯ͠ೖྗ͞ΕͨτʔΫϯΛ ը૾τʔΫϯͷઌ಄ʹՃ • ͜ΕΒͷτʔΫϯը૾τʔΫϯͱಉ༷ʹѻΘΕɺ 7J5ʣʹ͓͚ΔDMTτʔΫϯͱࣅׂͨΛ࣋ͭ
"QQFOEJYɿ$SPTT"UUFOUJPOCMPDL • 4FMG"UUFOUJPOϒϩοΫͷޙʹ$SPTT"UUFOUJPOΛ Ճͨ͠ઃܭ • <7BTXBOJ /*14>-%.ͱྨࣅͨ͠ΞʔΩςΫνϟ
"QQFOEJYɿ%J5CMPDLEFTJHO • %J59-Ϟσϧʹ͓͍ͯBEB-/;FSPΛ༻͍ͨ ߹͕࠷গͳ͍ܭࢉࢿݯͰ࠷ྑ͍ '*%,είΞΛୡ
"QQFOEJYɿ7JTJPO5SBOTGPSNFS <%PTPWJUTLJZ *$-3>
"QQFOEJYɿ*ODFQUJPO4DPSF *4 • *NBHF/FUͰࣄલֶशࡁΈͷ*ODFQUJPOOFUXPSLΛ༻͍ͨධՁࢦඪ • *ODFQUJPOOFUXPSL͕ࣝผ͘͢͠ɼࣝผ͞ΕΔϥϕϧͷଟ༷ੑ͕͋Δ΄Ͳ େ͖͘ͳΔࢦඪ <4[FHFEZ $713>
"QQFOEJYɿ'SFDIFU*ODFQUJPO%JTUBODF '*% • *NBHF/FUͰࣄલֶशࡁΈͷ*ODFQUJPOOFUXPSLΛ༻͍ͨධՁࢦඪ • ੜ͞Εͨը૾ͷಛ͕(5ը૾ͷಛͱͲͷఔ ࣅ͍ͯΔ͔ΛධՁ͢Δࢦඪ • '*%͕খ͍͞΄Ͳੜ͞Εͨը૾ͷ࣭͕(5ը૾ʹ͍ۙͱߟ͑ΒΕΔ
"QQFOEJYɿ1SFDJTJPO3FDBMM • *NBHF/FUͰࣄલֶशࡁΈͷ7((<4JNPOZBO *$-3>Λ༻͍ͯ ಛϕΫτϧू߹ΛಘΔ
"QQFOEJYɿ(GMPQT • 'MPQTɿුಈখԋࢉͷճ • (GMPQT 'MPQT • ը૾ੜλεΫͰΞʔΩςΫνϟͷෳࡶ͞ΛධՁ͢ΔࡍύϥϝʔλΛ༻͍Δͷ ͕Ұൠత
• ੑೳʹେ͖͘Өڹ͢Δը૾ղ૾ΛҰߟྀ͍ͯ͠ͳ͍ • Ϟσϧͷෳࡶ͞Λද͢ࢦඪͱͯ͠ෆेͳ߹͕͋Δ
"QQFOEJYɿఆྔత݁Ռ
"QQFOEJY(GMPQTͱ'*%ͷ૬ؔ • ΑΓଟ͘ͷ(GMPQTΛͭϞσϧ'*%͕͘ͳΔ
"QQFOEJYɿϞσϧαΠζͱύοναΠζͷݕ౼