Upgrade to Pro — share decks privately, control downloads, hide ads and more …

機械翻訳における文型パタンの部分的利用

Avatar for 自然言語処理研究室 自然言語処理研究室
March 31, 2006
Research
0
74

 機械翻訳における文型パタンの部分的利用

松田 聡史. 機械翻訳における文型パタンの部分的利用. 長岡技術科学大学課題研究報告書 (2006.3)
Avatar for 自然言語処理研究室

自然言語処理研究室

March 31, 2006
Tweet

More Decks by 自然言語処理研究室

See All by 自然言語処理研究室
データサイエンス14_システム.pdf
Avatar for 自然言語処理研究室 jnlp
0
380
データサイエンス13_解析.pdf
Avatar for 自然言語処理研究室 jnlp
0
480
データサイエンス12_分類.pdf
Avatar for 自然言語処理研究室 jnlp
0
330
データサイエンス11_前処理.pdf
Avatar for 自然言語処理研究室 jnlp
0
460
Recurrent neural network based language model
Avatar for 自然言語処理研究室 jnlp
0
130
自然言語処理研究室 研究概要(2012年)
Avatar for 自然言語処理研究室 jnlp
0
130
自然言語処理研究室 研究概要(2013年)
Avatar for 自然言語処理研究室 jnlp
0
93
自然言語処理研究室 研究概要(2014年)
Avatar for 自然言語処理研究室 jnlp
0
110
自然言語処理研究室 研究概要(2015年)
Avatar for 自然言語処理研究室 jnlp
0
180

Other Decks in Research

See All in Research
(NULLCON Goa 2025)Windows Keylogger Detection: Targeting Past and Present Keylogging Techniques
Avatar for Asuka Nakajima asuna_jp
2
550
[CV勉強会@関東 CVPR2025] VLM自動運転model S4-Driver
Avatar for Shin-kyoto shinkyoto
2
340
SkySense : A Multi-Modal Remote Sensing Foundation Model Towards Universal Interpretation for Earth Observation Imagery
Avatar for SatAI.challenge satai
3
260
とあるSREの博士「過程」 / A Certain SRE’s Ph.D. Journey
Avatar for Yuuki Tsubouchi (yuuk1) yuukit
6
3.3k
Delta Airlines® Customer Care in the U.S.: How to Reach Them Now
Avatar for booking.comcustomerservice bookingcomcustomersupportusa
PRO
0
100
[輪講] SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features
Avatar for Naoki Kato nk35jk
2
600
When Submarine Cables Go Dark: Examining the Web Services Resilience Amid Global Internet Disruptions
Avatar for Irvin Chen irvin
0
240
Creation and environmental applications of 15-year daily inundation and vegetation maps for Siberia by integrating satellite and meteorological datasets
Avatar for SatAI.challenge satai
3
140
Self-supervised audiovisual representation learning for remote sensing data
Avatar for SatAI.challenge satai
3
230
2025/7/5 応用音響研究会招待講演＠北海道大学
Avatar for Takuma OKAMOTO takuma_okamoto
1
110
大規模な2値整数計画問題に対する 効率的な重み付き局所探索法
Avatar for MIKIO KUBO mickey_kubo
1
280
引力・斥力を制御可能なランダム部分集合の確率分布
Avatar for Takahiro Kawashima wasyro
0
200

Featured

See All Featured
Dealing with People You Can't Stand - Big Design 2015
Avatar for Cassini Nazir cassininazir
367
26k
Designing for Performance
Avatar for Lara Hogan lara
610
69k
Speed Design
Avatar for Sergey Chernyshev sergeychernyshev
32
1k
Designing Dashboards & Data Visualisations in Web Apps
Avatar for Des Traynor destraynor
231
53k
Stop Working from a Prison Cell
Avatar for Chris Michel hatefulcrawdad
271
21k
The Success of Rails: Ensuring Growth for the Next 100 Years
Avatar for Eileen M. Uchitelle eileencodes
45
7.5k
Bootstrapping a Software Product
Avatar for Garrett Dimon garrettdimon
PRO
307
110k
Connecting the Dots Between Site Speed, User Experience & Your Business [WebExpo 2025]
Avatar for Tammy Everts tammyeverts
8
340
Let's Do A Bunch of Simple Stuff to Make Websites Faster
Avatar for Chris Coyier chriscoyier
507
140k
Embracing the Ebb and Flow
Avatar for Simon Collison colly
86
4.8k
Understanding Cognitive Biases in Performance Measurement
Avatar for Philip Tellis bluesmoon
29
1.8k
The Power of CSS Pseudo Elements
Avatar for Geoffrey Crofte geoffreycrofte
77
5.9k

Transcript

  1. ػց຋༁ʹ͓͚Δจܕɹɹ ɹ ɹ ɹɹɹύλϯͷ෦෼తར༻ ௕Ԭٕज़Պֶେֶɹిؾܥ ࢁຊݚڀࣨ ใࠂऀ দాɹ૱࢙ ࢦಋڭһ ࢁຊɹ࿨ӳ

    ॿڭत ฏ੒18೥ 2݄ 24೔

  2. 1.ݚڀഎܠ ύλϯ຋༁ͷ֓ཁ ɹ◆จ๏ɺ׳༻දݱΛݩʹύλϯΛ४උ ɹN1͕N2Λԉॿ͢Δ ɹN1 help N2 ɹ◆຋༁Λߦ͏จʹରͯ͠ύλϯΛ౰ͯ͸ Ίɺ༁จͷग़ྗΛߦ͏ɻ

  3. 1.ݚڀഎܠ ର༁ύλϯ ɹN1͕N2Λԉॿ͢Δ ɹN1 help N2 ೖྗจ ɹI help you

    ग़ྗจ ɹΘ͕ͨ͋͠ͳͨΛԉॿ͢Δ

  4. 1.ݚڀഎܠ ◆ ύλϯ຋༁ ɹ-ॏจɾෳจ͸ۤख ◆ ۙ೥ͷύλϯ຋༁ ɹ- ॏจɾෳจΛ໨తͱͨ͠ύλϯ࡞੒ ɹ- ύλϯ਺ͷ૿Ճ

    ɹ- ύλϯ࢖༻ޮ཰ͷ௿Լ →ύλϯΛ෦෼తʹར༻͢Δ͜ͱ͸ग़དྷͳ͍͔

  5. 2.໨త ෦෼తར༻Մೳͳύλϯͷ࡞੒ ɹ◆ύλϯར༻ޮ཰ͷ޲্ ɹ◆຋༁ݴޠ΁ͷॊೈͳରԠ ɹ◆຋༁ઌݴޠͰͷදݱͷ෯ͷ૿Ճ

  6. 3.CRESTύλϯ CRESTύλϯʹ͍ͭͯ ◆ॏจ΍ෳจͷ຋༁Λ໨తͱͯ͠࡞੒ ◆୯ޠϨϕϧͰ໿12ສύλϯ ◆ઢܗ෦෼ͱඇઢܗ෦෼͔ΒͳΔ

  7. 3.CRESTύλϯ ◆ ઢܗ෦෼ ɹ- ଞͷཁૉʹஔ͖׵͑ͯ΋શମͷҙຯ͕ มԽ͠ͳ͍ ◆ ඇઢܗ෦෼ ɹ- ஔ͖׵͑Δͱશମͷҙຯ͕มԽ͢Δ

    ɹN1͕N2Λԉॿ͢Δ ɹN1 help N2

  8. 4.ؔ࿈ݚڀ ◆ਆ໺Β[2005]ɹίʔύεΛݩʹ਺୯ޠఔ ౓ͷ໊ࢺ۟຋༁Λ໨తͱͨ͠ύλϯ࡞੒ ◆നҪΒ[2003]ɹίʔύεΛݩʹ୹จͷ຋ ༁Λ໨తͱͨ͠ύλϯΛ࡞੒ طଘͷύλϯ͔ΒύλϯΛ࡞੒͢Δͱ͍͏ ݚڀ͸ଘࡏ͠ͳ͍

  9. 5.ఏҊख๏ ύλϯΛ౰ͯ͸ΊΔࡍʹରԠ෦෼Λ୳͢ͷ Ͱ͸ඇޮ཰ ɹ→͋Β͔͡Ίύλϯͷ෼ׂΛߦ͏ ɹ→ରԠ෦෼Λࣗಈతʹ୳͢ख๏͕ඞཁ N1͕N2ʹͳ͔ͬͯΒN3ΛV4^meireiɻ V4 N3 after N1

    turn N2. N1͕N2ʹͳ͔ͬͯΒ after N1 turn N2.

  10. 5.ఏҊख๏ ◆ύλϯ͔Βඇઢܗ෦෼Λநग़ N1͕N2ʹͳ͔ͬͯΒN3ΛV4^meireiɻ V4 N3 after N1 turn N2.

  11. 5.ఏҊख๏ ύλϯઌ಄·ͨ͸຤ඌ-֤ඇઢܗ෦෼ Λύλϯީิͱͯ͠நग़ N1͕N2ʹͳ͔ͬͯΒ ʹͳ͔ͬͯΒN3ΛV4^meireiɻ V4 N3 after ɹafter N1

    turn N2. V4 N3 after N1 turn ɹturn N2.

  12. 5.ఏҊख๏ ؚ·ΕΔઢܗ෦෼͕ಉ͡ύλϯಉ࢜Λ෦෼ ύλϯͱͯ͠ొ࿥ N1͕N2ʹͳ͔ͬͯΒ after N1 turn N2. ʹͳ͔ͬͯΒN3ΛV4^meireiɻ V4

    N3 after

  13. 5.ఏҊख๏ લஔࢺɺؔ܎୅໊ࢺΛݩʹ෼ׂͨ͠ύλϯ ΋࡞੒ લஔࢺɹɹɹN1͸N2ʹV3^rentai຦ͷؒ ɹɹɹɹɹɹuntil N1 V3.past N2 ؔ܎୅໊ࢺɹN1͸N2Ͱग़͔͚͍ͯΔؒ ɹɹɹɹɹɹwhile

    N1 be on N2

  14. 6.ධՁ࣮ݧ 1. ɹ࡞੒ͨ͠ύλϯʹΑΔϚον਺ 2. ɹ࡞੒ͨ͠ύλϯͷҙຯతͳରԠ

  15. 6.ධՁ࣮ݧ1 CRESTύλϯͷݪจ͔Β ɹ◆ϥϯμϜʹ100จ ɹ◆ؔ܎୅໊ࢺͰ࢝·Δ100จ ɹ◆લஔࢺͰ࢝·Δ100จ Λநग़ 4छྨͷύλϯΛ౰ͯ͸ΊΔ

  16. 6.ධՁ࣮ݧ1(݁Ռ)         

               ࣮ݧจ*% Ϛον਺ ؔ܎୅໊ࢺ ඇม਺෦෼ $3&45ύλϯ લஔࢺ

  17. 6.ධՁ࣮ݧ1(݁Ռ) ֤จʹର͢ΔฏۉϚον਺ CRESTύλϯ ඇઢܗ෦෼ લஔࢺ ؔ܎୅໊ࢺ 16.6 100.7 0.6 196.7

  18. 6.ධՁ࣮ݧ2 ࡞੒ͨ͠ӳޠύλϯͱ೔ຊޠύλϯͰ ҙຯతͳରԠ͕ͱΕ͍ͯΔ͔ ˓ɿҙຯతͳରԠ͕ͱΕ͍ͯΔ ˚ɿӳ೔ɺ೔ӳ͍ͣΕ͔ʹؔͯ͠ɺҙຯͷ औΕͳ͍ඇઢܗ෦෼͕ଘࡏ͢Δ ×ɿҙຯతͳରԠ͕ͱΕ͍ͯͳ͍

  19. 6.ධՁ࣮ݧ2(݁Ռ) ύλϯͷҙຯతରԠ ˓ ˚ × ඇઢܗ෦෼ 43 47 10 લஔࢺ

    30 43 27 ؔ܎୅໊ࢺ 35 58 7

  20. 6.ධՁ࣮ݧ2(݁Ռ) ˓ɹN1 have a number of N2, ɹɹN1͸N2͕୔ࢁ͍ΔͷͰ ˚ɹIt is

    said among N2 in N1 that ɹɹN1ͷN2ͷؒͰ͸ ×ɹto V4 N3 ɹɹʹઌۦ͚ͯN3ΛV4.kako

  21. 7.ߟ࡯ ◆Ϛον਺ͷ૿Ճ ɹ→຋༁ઌݴޠʹ͓͚Δදݱͷ޿͕Γ͔͍ ◆Ϛονͨ͠ύλϯʹ͍ͭͯ ɹ→຋༁ͷਖ਼ղͱͳΔύλϯ͸ଘࡏ͢Δ ɹ→୹จͷ຋༁ʹ͸࢖༻Մೳ ◆׬શͳҙຯରԠͷऔΕͳ͍ύλϯ ɹ→ഉআ͢Δख๏ͷݕ౼͕ඞཁ

  22. 8.·ͱΊ CRESTύλϯΛݩʹ෦෼తύλϯΛ࡞੒ ඇઢܗ෦෼(લஔࢺɺؔ܎୅໊ࢺ)Λݩʹύ λϯΛ෼ׂɾ࠶ߏ੒ ಛఆͷจܕʹ͍ͭͯ͸ύλϯϚον਺͕େ ෯ʹ૿Ճ ҙຯͷऔΕͳ͍ύλϯͷଘࡏ(ࠓޙͷ՝୊)