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
DRL 組み合わせ最適化
Search
newzy
November 24, 2021
Research
0
83
DRL 組み合わせ最適化
newzy
November 24, 2021
Tweet
Share
Other Decks in Research
See All in Research
言語と数理の交差点:テキストの埋め込みと構造のモデル化 (IBIS 2024 チュートリアル)
yukiar
5
1.1k
EBPMにおける生成AI活用について
daimoriwaki
0
280
Human-Informed Machine Learning Models and Interactions
hiromu1996
2
580
CUNY DHI_Lightning Talks_2024
digitalfellow
0
500
Tiaccoon: コンテナネットワークにおいて複数トランスポート方式で統一的なアクセス制御
hiroyaonoe
0
430
Composed image retrieval for remote sensing
satai
3
240
PhD Defence: Considering Temporal and Contextual Information for Lexical Semantic Change Detection
a1da4
0
120
[輪講] Transformer Layers as Painters
nk35jk
4
690
資産間の相関関係を頑健に評価する指標を用いたファクターアローケーション戦略の構築
nomamist
0
140
Bluesky Game Dev
trezy
0
150
DeepSeek-R1の論文から読み解く背景技術
personabb
3
320
Weekly AI Agents News! 1月号 アーカイブ
masatoto
1
180
Featured
See All Featured
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
27
1.9k
Agile that works and the tools we love
rasmusluckow
328
21k
Java REST API Framework Comparison - PWX 2021
mraible
29
8.4k
I Don’t Have Time: Getting Over the Fear to Launch Your Podcast
jcasabona
32
2.1k
JavaScript: Past, Present, and Future - NDC Porto 2020
reverentgeek
47
5.2k
Visualization
eitanlees
146
15k
4 Signs Your Business is Dying
shpigford
182
22k
Music & Morning Musume
bryan
46
6.4k
GraphQLの誤解/rethinking-graphql
sonatard
68
10k
Learning to Love Humans: Emotional Interface Design
aarron
273
40k
Embracing the Ebb and Flow
colly
84
4.6k
Rebuilding a faster, lazier Slack
samanthasiow
80
8.9k
Transcript
POMO: Policy Optimization with Multiple Optima for Reinforcement Learning Kwon,
Yeong-Dae, et al. NeurIPS, 2020, vol.33
ཁ •Έ߹Θͤ࠷దԽʹ͓͚ΔɼਂڧԽֶश ͰͷFOEUPFOEͷۙࣅղ๏ɽ •طଘͷਂڧԽֶशख๏ͱൺֱͯ͠ɼ ܭࢉ࣌ؒɾਫ਼ͱʹେ͖͘վળͨ͠ •८ճηʔϧεϚϯͳͲͰݕূɽ 2/26
ಋೖ
Έ߹Θͤ࠷దԽ •८ճηʔϧεϚϯૹܭըɼφοϓβοΫ ͳͲʹද͞ΕΔΑ͏ͳ࠷దͳΈ߹ΘͤΛٻΊΔɽ 4/26 精度 計算時間 厳密解法 最適 遅い 近似解法
最適に 近い 早い https://onl.tw/vzkASMX
ڧԽֶशʢ3FJOGPSDFNFOU-FBSOJOH3-ʣ •3-ɿஞ࣍తͳҙࢥܾఆΛղ͘ख๏ɽ ྦྷੵใु͕࠷େʹͳΔΑ͏ͳํࡦΛݟ͚ͭΔ͜ͱ͕తɽ 5/26 ઃఆͱͯ͠ɼঢ়ଶू߹ɼߦಈू߹ɼใुؔΛ ઃఆ͢Δඞཁ͕͋Δɽ https://onl.tw/98fQVvW
ํࡦϕʔεͷ3&*/'03$& 6/26 •ํࡦ 𝜋 𝑠 ɿঢ়ଶ𝑠ʹ͓͚Δߦಈ𝑎Λग़ྗ͢Δؔ •𝜋! ɿύϥϝʔλ 𝜃ͰύϥϝʔλԽ͞Εͨํࡦ •ํࡦͷߋ৽ࣜɿ𝛼ֶशɼ𝐽
𝜋! తؔ 𝜃 ← 𝜃 + 𝛼∇! 𝐽 𝜋! •ํࡦޯͷࣜɿ𝔼ظɼ𝑅" ऩӹɼ𝑏 𝑠 ϕʔεϥΠϯ ∇! 𝐽 𝜋! = 𝔼#! ∇! log 𝜋! ⋅ 𝑅" − 𝑏 𝑠
ઌߦݚڀ
1PJOUFS/FUXPSLTʢʣ Έ߹Θͤ࠷దԽͰར༻͢ΔωοτϫʔΫ •ॏෳͳ͘બ͠ɼग़ྗύλʔϯྻΛੜ͢Δɽ •ೖྗใ͔Βಛநग़Λߦ͏FODPEFSͱɼFODPEFS ͷग़ྗΛར༻ͯ͑͠ͱͳΔܦ࿏Λग़ྗ͢ΔEFDPEFS͔ ΒͳΔɽ •FODPEFSͱEFDPEFSʹ-45.Λ༻ɽ 8/26
"UUFOUJPO .PEFMʢʣ 1PJOUFS/FUXPSLTͷվྑ൛ •1PJOUFS/FUXPSLTಉ༷ɼ&ODPEFSͱ%FDPEFSΛ༻͢Δ Ϟσϧɽ •-45.ഇࢭ͠ɼ.VMUJIFBE"UUFOUJPOΛ࠾༻ɽ 9/26
ख๏
ຊจͷख๏ͷΞΠσΞ 11/26 ࠷ॳͷߦಈɼޙͷΤʔδΣϯτͷߦಈʹେ͖͘ӨڹΛ༩͑Δɽ Έ߹Θͤ࠷దԽʹΑ͘ݟΒΕΔରশੑΛར༻ɽ
10.0 •3&*/'03$&XJUI#BTFMJOFɿయܕతͳํࡦޯϕʔεͷ 3-ΞϧΰϦζϜΛ༻ɽ •ෳͷҟͳΔ։࢝ߦಈΛࢦఆ͠ɼෳͷߦಈܥྻʢيಓʣ ΛಘΔɽ •ʻ45"35ʼτʔΫϯΛ༻͍ͳ͍ɽ 12/26 従来 POMO
10.0 ∇! 𝐽 𝜃 ≈ 1 𝑁 6 $%& '
𝑅 𝜏$ − 𝑏$ 𝑠 ∇! log 𝑝! 𝜏$ ∣ 𝑠 𝑤ℎ𝑒𝑟𝑒 𝑝! 𝝉$ ∣ 𝑠 ≡ @ "%( ) 𝑝! 𝑎" $ ∣ 𝑠, 𝑎&:"+& $ يಓ 𝝉$ = 𝑎& $ , 𝑎( $ , … , 𝑎) $ GPS 𝑖 = 1,2, … , 𝑁 ڞ༗ϕʔεϥΠϯ 𝑏$(𝑠) = 𝑏TIBSFE (𝑠) = 1 𝑁 6 ,%& ' 𝑅 𝝉, GPS 𝑖 = 1,2, … , 𝑁 13/26
܇࿅෦ͷٖࣅίʔυ 14/26
*OTUBODF"VHNFOUBUJPOɿਪख๏ •ը૾ॲཧͷσʔλΦʔάϝϯςʔγϣϯ͔Βணɽ •ࠓճ͏࠲ඪɼYͷ୯Ґਖ਼ํܗʢୈҰݶʣͷ ͷΛར༻ɽ 15/26 今回使う Instance Augmentation
ਪ෦ͷٖࣅίʔυ 16/26
࣮ݧ
࣮ݧ ࣮ݧ༰ •10.0Λ༻͍ͯɼҎԼͷΛղ͍ͨ݁ՌΛଞͷදతख๏ͱ ൺֱɽ ८ճηʔϧεϚϯ ༰ྔ੍͋Γͷૹܭը φοϓβοΫ
18/26
ֶशۂઢɿ८ճηʔϧεϚϯ 19/26 50地点 100地点
८ճηʔϧεϚϯʢ541ʣ 20/26
८ճηʔϧεϚϯʢ541ʣ 21/26
༰ྔ੍͋Γͷૹܭըʢ$731ʣ 22/26
φοϓβοΫʢ,1ʣ 23/26
࣮ݧͷ·ͱΊ •ҟͳΔઃఆͷͭͷΈ߹Θͤ࠷దԽʹରͯ͠ɼ ಉҰͷ܇࿅ख๏ͱ//ΞʔΩςΫνϟΛ༻͍ͯ༗ͳ݁ՌΛ ಘͨɽ •܇࿅ɾਪख๏ͱͯ͠ͷ10.0ɼਪख๏ͱͯ͠ͷ *OTUBODF"VHNFOUBUJPOͲͪΒޮՌతͳख๏Ͱ͋Δ͜ͱ Λ֬ೝͨ͠ɽ 24/26
·ͱΊ ຊจͰΈ߹Θͤ࠷దԽʹ͓͍ͯɼରশੑΛར༻ ͯ͠3-ͷαϯϓϧޮਫ਼ ਪ࣌ؒΛॖ͢Δख๏Λ հͨ͠ɽ 25/26
ࢀߟจݙ ,XPO :FPOH%BF FUBM10.01PMJDZ0QUJNJ[BUJPOXJUI .VMUJQMF0QUJNBGPS3FJOGPSDFNFOU-FBSOJOH "EWBODFTJO /FVSBM*OGPSNBUJPO1SPDFTTJOH4ZTUFNT
,PPM 8PVUFS )FSLF WBO)PPG BOE.BY8FMMJOH"UUFOUJPO -FBSOUP4PMWF3PVUJOH1SPCMFNT *OUFSOBUJPOBM$POGFSFODF PO-FBSOJOH3FQSFTFOUBUJPOT 7JOZBMT 0SJPM .FJSF 'PSUVOBUP BOE/BWEFFQ+BJUMZ1PJOUFS /FUXPSLT "EWBODFTJO/FVSBM*OGPSNBUJPO1SPDFTTJOH 4ZTUFNT 26/26