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
A Quantum Computational Approach to Corresponde...
Search
peisuke
July 04, 2020
Technology
0
650
A Quantum Computational Approach to Correspondence Problems on Point Sets
全日本CV勉強会
CVPR2020読み会発表資料
peisuke
July 04, 2020
Tweet
Share
More Decks by peisuke
See All by peisuke
LangGraphで始めるマルチエージェントシステム
peisuke
13
3.7k
AutoGenで作るLLM Agen
peisuke
1
1.2k
Self-RAG: Learning to Retrieve, Generate and Critique through Self-Reflections
peisuke
9
1.4k
Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields
peisuke
0
7.9k
LangChain Toolsの運用と改善
peisuke
4
2.5k
GNeRF: GAN-based Neural Radiance Field without Posed Camera
peisuke
1
690
TTS Skins: Speaker Conversion via ASR
peisuke
0
360
F0-Consistent Many-to-many Non-parallel Voice Conversion via Conditional Autoencoder
peisuke
0
160
YOLACT real-time instance segmentation
peisuke
1
230
Other Decks in Technology
See All in Technology
リクルートのデータマネジメント組織に 求められてきたコト
recruitengineers
PRO
4
340
いまからでも遅くない! コンテナでWebアプリケーションを 動かしてみよう(1)Gitハンズオン
nomu
0
280
20240906_JAWS_Yamanashi_#1_leap_beyond_the_AWS_all_certifications
tsumita
1
100
【Λ(らむだ)最近のアプデ情報 / RPALT20240904
lambda
0
160
音声AIエージェントの世界とRetell AI入門 / Introduction to the World of Voice AI Agents and Retell AI
rkaga
3
470
Binary Authorizationと友達になろう / Let's be friends with Binary Authorization
iselegant
2
130
FastlyとfalcoでNode.jsレスな Webサーバー構築 : IPTV版 ABEMAアプリインフラ刷新 / Building web server using Fastly and falco without Node.js for IPTV ABEMA app
nodaguti
0
240
Estrategias de escalabilidade para projetos web
jessilyneh
1
150
20240828_EM_Fest
kzkmaeda
4
360
[RSJ24] Object Retrieval in Large-Scale Indoor Environments Using Dense Text with a Multi-Modal Large Language Model
keio_smilab
PRO
0
220
React Aria で実現する次世代のアクセシビリティ
ryo_manba
3
290
すぐに始めるAWSコスト削減。短期でできる改善策と長期的な運用負荷軽減への取り組み方を解説
ncdc
1
120
Featured
See All Featured
Building an army of robots
kneath
302
42k
Why You Should Never Use an ORM
jnunemaker
PRO
53
8.9k
Fantastic passwords and where to find them - at NoRuKo
philnash
47
2.7k
What's new in Ruby 2.0
geeforr
340
31k
Imperfection Machines: The Place of Print at Facebook
scottboms
263
13k
[Rails World 2023 - Day 1 Closing Keynote] - The Magic of Rails
eileencodes
27
1.6k
The Invisible Customer
myddelton
119
13k
Scaling GitHub
holman
458
140k
Web development in the modern age
philhawksworth
204
10k
The Pragmatic Product Professional
lauravandoore
30
6.2k
GraphQLとの向き合い方2022年版
quramy
43
13k
Into the Great Unknown - MozCon
thekraken
28
1.4k
Transcript
"2VBOUVN$PNQVUBUJPOBM"QQSPBDIUP $PSSFTQPOEFODF1SPCMFNTPO1PJOU4FUT "VUIPST7(PMZBOJL $5IFPCBMU 1SFTFOUFS!QFJTVLF
荈䊹稱➜ せ谏劤 価➜ 䨽㾩"#&+" 灇瑔؝ٝؾُ٦ةؽآّٝծٗنحز 崞⹛ 5XJUUFS!QFJTVLF (JUIVC IUUQTHJUIVCDPNQFJTVLF 2JJUB
IUUQTRJJUBDPNQFJTVLF 4MJEF4IBSFIUUQTXXXTMJEFTIBSFOFU'VKJNPUP,FJTVLF
嚊銲 • ة؎زٕ • "2VBOUVN$PNQVUBUJPOBM"QQSPBDIUP$PSSFTQPOEFODF1SPCMFNT PO1PJOU4FUT • וז锷俑 • ꆀ㶨،ص٦ؚٔٝوءٝד挿纇窟さׅ倯岀חאְג
• 䊨㣗挿כ • 㔐鯄遤ך䱱稊26#0ד邌植׃鵚⡂涸ח鍑ְ !! ∈ 0, 1 26#0
ꆀ㶨،ص٦ؚٔٝوءٝך铡僇 • ⡦ָ⳿勻ַ • 2VBESBUJDVODPOTUSBJOFECJOBSZPQUJNJ[BUJPO剑黝⻉ • Rחꟼׅ穈さׇ剑黝⻉㉏겗 !! ∈ 0,
1 • 26#0ה׃ג鍑ֽ㉏겗ך⢽ • .BY$VU • ؚٓؿ䕙葿㉏겗 • 1BSUJUJPO1SPCMFN
ꆀ㶨،ص٦ؚٔٝوءٝך铡僇 • ⹛⡲⾱椚ך嚊銲 • ؎آٌؚٝرָٕ暟椚涸ח㹋鄲ׁגְ • 㛇䎿朐䡾חזהֹך朐䡾錁庠ׅה剑黝鍑䖤
ꆀ㶨،ص٦ؚٔٝوءٝך铡僇 • ꆀ㶨ؽحزהכ • אך朐䡾⢽ִלהָꅾזさֲ • ⢽ִלծ铣《הך然桦דָծ ך然桦דָ⳿⸂ׁ • ך朐䡾
⟩ |0 ծך朐䡾 ⟩ |1 ה邌׃ծ ـٗحم椔 ⟩ |# = ⟩ %|0 + ⟩ (|1 ך״ֲחꅾיさׇ邌ׅ %ծ &כ醱稆侧ծ % ! + & !弫ׅ • ـٗحم椔〸㔳כꆀ㶨ؽحزך朐䡾ך荈 歋䏝㔳爙׃ך
ꆀ㶨،ص٦ؚٔٝوءٝך铡僇 • ءُٖ٦ر؍ٝؖ٦倯玎䒭 • ꆀ㶨ؽحزך朐䡾ך儗涪㾜ך倯玎䒭 • 麊⹛倯玎䒭ךꆀ㶨غ٦آّٝך״ֲזךծ岚ך朐䡾䗍ⴓׅ✲ד実ת • 植㖈ך朐䡾חعىٕزص،ٝ)⡲欽ׇֿׁהד如ך朐䡾邌ׇ •
عىٕزص،ٝ • 朐䡾ךؒطؘٕ٦爙ׅ • 〢Ⱙ暟椚חֶֽ⸇鸞䏝ְז؎ً٦آ • 禸ָ㹀䌢朐䡾חזגְהծ)ď&ď弫׃朐䡾הז • عىٕزص،ٝך剑㼭㔿剣⦼ח㼎䘔ׅ㔿剣ꟼ侧ָꆀ㶨ؽحزך㛇䎿 朐䡾ծָֿꆀ㶨،ص٦ؚٔٝחֶֽ剑黝鍑ח㼎䘔
ꆀ㶨،ص٦ؚٔٝوءٝך铡僇 • ػؐٔ遤 • ꆀ㶨ؽحز怴皾ך剑㛇劤涸ז怴皾 • ػؐٔ遤穈さׇֿהדعىٕزص،ٝ圓眠 • " !
+ $ !弫׃תתד㢌䳔דֹ • ꆀ㶨،ص٦ؚٔٝوءٝחֶֽعىٕزص،ٝ • ꆀ㶨،ص٦ؚٔٝوءٝדכِ٦ؠָعىٕزص،ٝך⦼鏣㹀 • عىٕزص،ָٝ؎آٌؚٝرٕהז״ֲחׅ • 㹀䌢朐䡾⡲ֿהדծ؎آٌؚٝرٕך剑黝鍑ָ実ת
26#0ַ؎آٌؚٝرٕפך㢌䳔 • 鍑ְֹ26#0 • ؎آٌؚٝرٕ • ⟃♴ך״ֲח㢌侧㢌䳔
ꆀ㶨،ص٦ؚٔٝوءٝתה • ꆀ㶨،ص٦ؚٔٝوءٝחֶֽعىٕزص،ٝה׃גծ؎آ ٌؚٝرٕ♷ִ • ؎آٌؚٝرٕך㛇䎿朐䡾錁庠ׅ✲ד剑黝鍑ָ䖤 • ؎آٌؚٝرٕכ26#0ה瘝⣣ד֮ծ穈さׇ㉏겗ך剑黝 ⻉㉏겗鍑ֻֿהָ⳿勻
ꆀ㶨،ص٦ؚٔٝ欽ְ㔐鯄ך鎘皾% • 㔐鯄遤ծְֻאַך㛇䎿חⴓ鍑׃ծך穈さׇח״ ג邌植 • 如⯋ך㔐鯄遤40 כծ⟃♴ך״ֲח剅ֽ • 3כ*
.ךꅾ➰ֹㄤה鋅זׇծ⟃♴ך״ֲז㛇䎿2ך穈 さׇ㉏겗הׅ
ꆀ㶨،ص٦ؚٔٝ欽ְ㔐鯄ך鎘皾% • ⯓玎ך穈さׇ㉏겗⢪ְծ挿纇Y Zך騃ꨄ剑㼭⻉ & ' = ) − '+
% = ) − ,& + − ,% + − ⋯ − ,' + % ״
ꆀ㶨،ص٦ؚٔٝ欽ְ㔐鯄ך鎘皾% • %הず圫ח㔐鯄遤ⴓ鍑ׅ .ך⛦ך갪חꟼׅ㛇䎿 .חꟼׅ㛇䎿
㢳⡤㉏겗ⵃ欽׃ؚٗ٦غٕ挿纇窟さ • 㸘㸾ַ挿纇פ • 僤א鏣縧ׅהծ♰剣䒷⸂ד䒷ַさג遼瑱ׅ • 遼瑱ׅ澓ָծ䒷ַさֲؒطؘٕ٦ָ剑㣐הז • 湱㼎㪦ָ㔿㹀ׁ僤㔚ה僤㔚鏣縧׃ծ✼ְָ䒷䓸さֲ ؒطؘٕ٦ָ剑㣐ְֹ朐䡾⡲
• 僤㔚挿纇ה鋅甧גגծ剑㣐ؒطؘٕ٦ד遼瑱׃ծ挿纇ָ窟さׁ 朐䡾הז ؒطؘٕ٦㼭 ؒطؘٕ٦㣐 ؒطؘٕ٦㣐
㢳⡤㉏겗 • 湱✼⡲欽ׅ⡤⟃♳ך禸ַז麊⹛実㉏겗 • 㸘㸾ד僤ず㡦ָ♰剣䒷⸂ד䒷ֹさֲꥷך䞨僤麊遤זו ( . R ٥٥٥ꅾ⸂⸇鸞䏝 ٥٥٥颵ꆀ
٥٥٥⡘縧
㢳⡤㉏겗挿纇窟さח䘔欽 <7(PMZBOJL > • Ⰻגך挿ず㡦㼎䘔➰ֽծؒطؘٕ٦剑㣐⻉ • 挿ָ♧荜׃㜥さח♧荜׃挿ず㡦ךؒطؘٕ٦כ̺חזծ ➭ך铎㼎䘔湱㼎涸ח搀鋔דֹ • *$1،ٕ؞ٔؤيך״ֲז㼎䘔➰ֽ㉏겗涪欰ׇׁ׆חծؚ
ٗ٦غٕז窟さָ〳腉 ė ٥٥٥ⴓ嫡ָחזךּך㼭ׁז⦼
挿纇窟さ䩛岀ך何葺晛 <7(PMZBOJL > • 㢳⡤㉏겗ךתת鍑ֻה♶㸜㹀ծ尶㿊ך 勲鎘皾ָ䗳銲 • /FHBUJWFFMFNFOUXJTFSFDJQSPDBMUSBOTGPSNPQFSBUPSⵃ欽 ׃גؒطؘٕ٦ꟼ侧㢌䳔 • Ⰻגך挿ず㡦דؙٔٝ䓸湫䱸剑黝⻉
• (װNזוךػًٓ٦ةכד葺ְ
醱侧挿纇ؙٔٝ׃ג如䕎䒭ח㢌䳔 • ⚛鹌Uכ*$1ず圫ח⚕挿纇⚥䗰⻉׃גꤐ ( ) = + ",$ ," −
).$ ! = + ",$ ," − /% .$ − /! .$ − ⋯ − /& .$ ! ぐղךYחאְגծ鵚⩸,⦐ךZח㼎׃גךؙٔٝ鎘皾 㹋欽♳Ⰻגך挿חؙٔٝ顦הꅾְ ⰋגךYחאְגךؙٔٝ⚛ץ 如䕎䒭חׅكؙزٕRך銲稆כַ Rך穈さׇד3皾⳿
㹋꿀勴⟝ • 傀㶷ך鎘皾堣דءىُٖ٦ءّٝ㹋꿀 • %8BWFחכ➙㔐ך㉏겗鍑ֽ堣腉כזְהךֿה • 㔐鯄遤ך铎䊴הծ挿纇窟さ铎䊴חאְג㜠デ • ر٦ةإحز •
GJTI RVCJU LBOKJ DPNQPTFS %ⵃ欽 • זֶ%כ鎘皾♶〳腉ךד 㹋꿀ז׃
窟さ㹋꿀 • 㹋꿀勴⟝ • ر٦ةإحزכGJTI挿 • ؙٔٝח欽ְ挿侧\ ^
• ⴱ劍،ٓ؎ًٝز铎䊴< Ģ>ך眔㔲ד㔐㹋꿀 • أ؛٦ٕろ荈歋䏝ך窟さ • 㹋꿀穠卓 • 㼎䘔挿ח欽ְ侧ָ⟃ⰻד֮ל 勲 • 挿侧㟓װׅהوحثؚٝדֹזַ • أ؛٦ٕろ窟さדכTJOHVMBSJUZUIFPSFNח״ 挿纇ָ挿ח簭鷌ׅ
㹋꿀穠卓 • ⴱ劍،ٓ؎ًٝز铎䊴ך㹋꿀 • וך錬䏝חꟼ׃ג,ָ⟃♴ד֮ ל㸜㹀׃ג窟さ〳腉 • ؚٗ٦غٕז窟さָ〳腉דֿ֮ה 爙㇗׃גְ •
ظ؎ؤחꟼׅ㹋꿀 • ظ؎ؤꆀ圫ղח鏣㹀 • ,ָ㼭ֽׁלظ؎ؤח㼎׃גٗغ أزז窟さָ〳腉
㹋꿀穠卓 • ألؙزٍؘٕحفך锃叨 • 剑㼭㔿剣⦼ה殢ח㼭ְׁ㔿剣⦼ךٍؘحفחאְג㹀䚍涸ח锃叨 • 鸐䌢ٍؘحفָ㼭ְׁקו 勲ָ♶㸜㹀 • 㹋꿀ד涪欰ׅألؙزٍؘٕحفָוך玎䏝㣐ֹֽל״ְַך鍑匿כ ꨇ׃ְָծ窟さ荈⡤כ♳䩛ֻ遤גְךד⼧ⴓזٍؘحفָ䖤גְ
ה罋ִג葺ְ
תה • ꆀ㶨،ص٦ؚٔٝⵃ欽׃挿纇窟さ • 㔐鯄遤穈さׇ剑黝⻉㉏겗ד鵚⡂ • 㢳⡤㉏겗欽ְ挿纇窟さך䩛岀ك٦أחꆀ㶨،ص٦ؚٔٝ ד窟さ遤 • 鸐䌢ך鎘皾堣⢪ְծ%ך挿纇㼎韋ח窟さ㹋꿀遤ְծؚ
ٗ٦غַٕאظ؎ؤח䓼ְ窟さָ〳腉ד֮✲爙׃