Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
ACL精神医療論文まとめ 8min LT
Search
uchi_k
November 02, 2019
Research
1.4k
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
ACL精神医療論文まとめ 8min LT
uchi_k
November 02, 2019
More Decks by uchi_k
See All by uchi_k
ACL2020 Category Survey: Sentiment Analysis
uchi_k
2
3.5k
抽出的文書要約における hetero graph の応用 Heterogeneous Graph Neural Networks for Extractive Document Summarization
uchi_k
0
1.3k
前処理が単語埋め込みに与える影響 A Comprehensive Analysis of Preprocessing for Word Representation Learning in Affective Tasks
uchi_k
2
1.1k
Graph Neural Networks のビジネス応用可能性 heterogeneous graph と論文再現性について
uchi_k
1
3.5k
【論文紹介】医用画像への転移学習の有効性について Transfusion: Understanding Transfer Learning for Medical Imaging
uchi_k
4
3.8k
Graph: A Survey of Graph Neural Networks, Embedding, Tasks and Applications
uchi_k
1
1.3k
Other Decks in Research
See All in Research
高性能計算機クラスタを用いた大規模点群処理による森林の単木抽出と構造解析
kentaitakura
1
140
Développer des solutions de réduction des émissions de méthane entérique : 1ers résultats Méthane 2030
institudelelevage
PRO
0
180
データサイエンティストの就労意識~2015 → 2026 一般(個人)会員アンケートより
datascientistsociety
PRO
0
930
RS-Agent: Automating Remote Sensing Tasks through Intelligent Agent
satai
3
580
ros2-perf-multihost: Automated Coordination Framework for Objective Architecture Evaluation in Distributed Systems
takasehideki
0
130
ふとした出会いで生まれたSkillが、 社内利用1位になるまで
mikimhk
22
24k
Visual SLAM未来予測 / Future Prediction in Visual SLAM
koide3
1
1.1k
GLIM とMegaParticles:正規分布近似の限界とタイトカップリング&パーティクルフィルタの進展 / GLIM and MegaParticles : Progress of the distribution representation in SLAM
koide3
0
900
某助成金プロジェクト採択に向けて企業研究所のアウトリーチ専任者がやったこと
afroscript
0
190
ros2-perf-multihost: 分散システムにおける客観的なアーキテクチャ評価フレームワーク
takasehideki
0
280
医療LLMの現在地〜最新研究から社会実装までを考える〜
kento1109
1
1.8k
SLAMはどこまで解決されたのか?
tomonom
0
1.4k
Featured
See All Featured
Paper Plane
katiecoart
PRO
4
53k
Easily Structure & Communicate Ideas using Wireframe
afnizarnur
194
17k
Become a Pro
speakerdeck
PRO
31
6.3k
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
31
2.9k
DBのスキルで生き残る技術 - AI時代におけるテーブル設計の勘所
soudai
PRO
68
58k
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
141
35k
Kristin Tynski - Automating Marketing Tasks With AI
techseoconnect
PRO
0
530
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
37
6.6k
JavaScript: Past, Present, and Future - NDC Porto 2020
reverentgeek
52
6.1k
WENDY [Excerpt]
tessaabrams
14
39k
Accessibility Awareness
sabderemane
1
220
Facilitating Awesome Meetings
lara
57
7.2k
Transcript
8min LT: What Makes a Good Counselor? Learning to Distinguish
between High-quality and Low-quality Counseling Conversations ΄͔
Uchihashi Kenshi uchi_k @wednesdaymuse About me •yuni, inc. •Freelance Machine
Learning Engineer / Researcher •xpaper.challenge ӡӦ •former ະ౿ ژେใӃ ੴҪݚڀࣨ
yuni, inc. Unity × Unique ɾਫ਼ਆͷෆௐΛධՁ͢Δͷ͕͍͠ ɾ΄ͱΜͲͷਓ͕࣏ྍʹߦ͘ඞཁ͕ͳ͍ͱզຫͯ͠͠·͏ ɾܧଓతͳ࣏ྍʹ͔͔Δҩྍඅ͕ඇৗʹߴֹ ͜ΕΒͷ՝ʹର͠σʔλʹΑΔղܾࡦΛఏࣔ͠ɺ ਫ਼ਆͷෆௐΛؾܰʹ૬ஊ͠ɺظతͳࢧԉ͕ड͚ΒΕΔ
ϓϥοτϑΥʔϜΛఏڙ͢Δ
#Abstracts 'JHVSBUJWF6TBHF%FUFDUJPOPG4ZNQUPN8PSETUP*NQSPWF 1FSTPOBM)FBMUI.FOUJPO%FUFDUJPO ༩͑ΒΕͨจষ͕ݸਓͷ݈߁ঢ়ଶʹؔ͢ΔݴٴͰ͋Δ͔൱͔Λఆ͢ Δ 8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTI CFUXFFO)JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT ΧϯηϦϯάʹ͓͚ΔձͰͷ૬ޓ࡞༻Λղੳ͢Δ͜ͱͰɺߴ࣭ ͳΧϯηϦϯάͷݴޠతಛΛಋ͖ɺΧϯηϦϯάͷ࣭ΛධՁ͢
ΔϞσϧΛ࡞ͨ͠ 0CTFSWJOH%JBMPHVFJO5IFSBQZ $BUFHPSJ[JOHBOE'PSFDBTUJOH#FIBWJPSBM$PEFT ରʹΑͬͯಋ͔ΕΔߦಈΛϥϕϦϯά͠ɺ༧ଌϞσϧΛ࡞͢Δ͜ ͱͰɺΧϯηϥʔʹϦΞϧλΠϜͰΨΠμϯεΛఏڙ͢Δ
#Abstracts 5PXBSET"VUPNBUJOH)FBMUIDBSF2VFTUJPO"OTXFSJOHJOB /PJTZ.VMUJMJOHVBM-PX3FTPVSDF4FUUJOH ೆΞϑϦΧʹ͓͚ΔνϟοτϔϧεέΞʹ͓͍ͯɺճͷ৴པੑ͕ ͍ʢεϖϧϛεɺ༰͕͓͔͍͠ͳͲʣ͕͋ͬͨͷͰɺςϯϓϨʔτ Λࣄલ࡞͔ͦ͜͠Βબͯ͠Β͏ܗࣜʹ͢Δ͜ͱͰԠͱਖ਼ ֬ੑ͕େ্͖ͨ͘͠ ˞લऀ̎ͭʹ͍ͭͯɺλεΫͷհϝΠϯͰബ͘͘հ͍͖ͯ͠·͢ɾɾɾʂ 'JOEJOH:PVS7PJDF5IF-JOHVJTUJD%FWFMPQNFOUPG.FOUBM )FBMUI$PVOTFMPST
ΧϯηϥʔͷաఔΛɺݸਓ͕ܦݧʹج͍ͮͯݴޠߦಈΛมԽ͞ ͤΔఔΛఆྔԽ͢Δ͜ͱʹΑͬͯͰ͖ΔΑ͏ʹ͠ɺൃలύλʔ ϯΛྨܕԽͨ͠
'JHVSBUJWF6TBHF%FUFDUJPOPG4ZNQUPN8PSET UP*NQSPWF1FSTPOBM)FBMUI.FOUJPO%FUFDUJPO #Personal Health Mention Detection "$- "EJUI*ZFS $4*30%BUB "EJUZB+PTIJ
4BSWOB[,BSJNJ 3PTT4QBSLT $FDJMF1BSJT DJUBUJPOT ݸਓͷ݈߁ঢ়ଶʹؔ͢ΔݴٴΛݕ͢ΔλεΫʢ1).EFUFDUJPOʣΛ $//CBTFEͳख๏Ͱղ͘͜ͱͰ4P5"Λୡ 1FSTPOBM)FBMUI.FOUJPO 1.) %FUFDUJPOͱɺຊ ਓ͘͠ਓ͕ঢ়݈߁ঢ়ଶΛܦݧ͍ͯ͠Δ͔Ͳ͏͔Λ ݕ͢ΔλεΫ ʮே͔Β֏͕ग़Δʯ1).͕ͩɺʮिؒҎ্ͷ֏͕Μ ͷةݥੑ͕͋Γ·͢ʯҰൠతͳ֏ʹؔ͢ΔใͰ͋ͬͯ 1).Ͱͳ͍ ݸਓͷ݈߁ঢ়ଶʹؔ͢ΔݴٴΛਖ਼֬ʹநग़͢Δ͜ͱͰɺපؾ ͷྲྀߦͷஹީҩྍਪનͳͲʹܨ͛ΒΕΔ
#Method #CNN 1JQFMJOFͱ'FBUVSF"VHNFOUBUJPOͷ̎ͭͷΞϓϩʔνΛఏҊɻ ൺᄻతදݱ͔൱͔Λݕ͢ΔϞδϡʔϧΛՃ͠ɺൺᄻతදݱͰͳ͍߹ ͷΈ1.)EFUFDUJPOΛߦ͏1JQFMJOFΞϓϩʔνͱɺϞδϡʔϧͷग़ྗ Λಛͱͯ͠ΈࠐΉ'FBUVSF"VHNFOUBUJPOΞϓϩʔνͷ̎ͭΛఏҊ 'JHVSBUJWF6TBHF%FUFDUJPOPG4ZNQUPN8PSETUP*NQSPWF1FSTPOBM)FBMUI.FOUJPO%FUFDUJPO pMUFST LFSOFMT
QPPMTJ[FESPQPVU CBUDITJ[F FQPDIT "EBNͷ$//͕ϕʔεϥΠϯ
#Result #Figure 1JQFMJOFͰੑೳ্͕ݟΒΕͳ͔͕ͬͨɺ 'FBU"VHΞϓϩʔνͰ'ͷ্͕ݟΒΕͨ 8PSEFNCFEEJOHख๏ΛछʢXPSEWFD (MP7F 8PSE/FU .F4)ͳͲʣࢼ͠ɺͦͷฏۉͰධՁɻ'FBU"VHͷΈͰ্ɻ 'JHVSBUJWF6TBHF%FUFDUJPOPG4ZNQUPN8PSETUP*NQSPWF1FSTPOBM)FBMUI.FOUJPO%FUFDUJPO
ප໊͝ͱͷධՁͰ΄ͱΜͲͷ߹'FBU"VH͕ϕʔεϥΠϯΛ্ճͬ ͍ͯΔ
#Conclusion ·ͱΊ ͦͦ1.)EFUFDUJPOͱ͍͏λεΫࣗମΛΒͳ͔͕ͬͨɺݸਓͷ ݈߁ঢ়ଶʹؔ͢ΔݴٴΛநग़Ͱ͖Δ͜ͱͰΞυςΫҬͷײછਐߦ ঢ়گɺපؾͷૣظൃݟͳͲ༷ʑͳԠ༻͕Ͱ͖ͦ͏ͩͳͱࢥͬͨ 'JHVSBUJWF6TBHF%FUFDUJPOPG4ZNQUPN8PSETUP*NQSPWF1FSTPOBM)FBMUI.FOUJPO%FUFDUJPO 1.)ͷಛੑͱͯ͠ɺҰൠతͳݴٴͰͳࣗ͋͘Δ͍ਓͷ݈߁ঢ়ଶ ʹ͍ͭͯͷݴٴͷΈநग़͠ͳ͚ΕͳΒͳ͍͕ɺ͜͜ʹൺᄻతදݱͰࣦ ഊ͢Δέʔε͕͋Δ͜ͱΛൃݟͨ͜͠ͱࣗମ໘ന͍ʢͦΕ͕ͲΕ͘Β͍ ˠʮύϦ͕͘͠ΌΈ͢ΔͱɺϤʔϩού෩अΛͻ͘ʯͱ͍͏ྫ͕ڍ͛
ΒΕ͍͕ͯͨɺຊޠͩͱ͍͍ྫ͕ࢥ͍͖ͭ·ͤΜͰͨ͠
8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTICFUXFFO )JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT #Conversation Analysis "$- 7FSPOJDB1FSF[3PTBT 6OJWFSTJUZPG.JDIJHBO FUBM ΧϯηϦϯάʹ͓͚ΔձͰͷ૬ޓ࡞༻Λղੳ͢Δ͜ͱͰɺߴ࣭ͳ
ΧϯηϦϯάͷݴޠతಛΛಋ͖ɺΧϯηϦϯάͷ࣭ΛධՁ͢ΔϞσϧΛ࡞ ΧϯηϦϯά໔ڐͳͲ͕ͳͯ͘ՄೳͳͨΊɺ࣭ͷԼ͕ࣾձతͳʹ ͳ͍ͬͯΔɻ࣭ͷߴ͍ɾ͍ΧϯηϦϯάͷಛΛಋ͖ɺͦΕΒΛϦΞϧλ ΠϜͰධՁͨ͠ΓΧϯηϥʔͷڭҭʹͭͳ͛Δ͜ͱͰ͜ͷΛղܾ͍ͨ͠ ΧϯηϦϯάྫɻ$͕ΫϥΠΞϯτͰ5͕Χϯηϥʔʢηϥϐετʣ
#Dataset #Motivation Interviewing #Labeling .PUJWBUJPOBM*OUFSWJFXJOH .* ͱݺΕΔɺΧϯηϦϯά͕ ߦΘΕ͍ͯΔσʔλͷΈΛ:PV5VCFͳͲ͔Βऔಘ͠ɺखಈͰϥϕϦϯά ΧϯηϦϯάͷ࣭ʹؔͯ͠ɺઌߦݚڀΛࢀߟʹධՁɻͬ͘͟Γݴ͏ ͱɺ
8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTICFUXFFO)JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT ʙఔͷΧϯηϦϯάಈըݸ͔Βɺݸͷߴ࣭ಈըͱ ͷ࣭ಈըʹϥϕϦϯά ձ༰ਫ਼ਆෆௐɺېԎɺΞϧίʔϧɺༀɺμΠΤοτͳͲͷ݈߁τϐο ͕த৺ ࣭ͷߴ͍ΧϯηϦϯάϦϑϨΫγϣϯϦεχϯάʢฉ͖ͯ͠૬खͷ ߟ͑Λཧ͢Δํ๏ʣɺ࣭ɺܾఆͷαϙʔτɺڠௐͳͲͷಛ͕͋Δ ࣭ͷ͍ΧϯηϦϯάϦεχϯά͕࣌ؒ͘ൃ͕͍࣌ؒɺΫϥΠ Ξϯτͱରཱ͢Δɺ૬खͷڐՄͳ͠ʹΞυόΠεΛ͢ΔͳͲͷಛ͕͋Δ
#Words, Turn Analysis #Question, Reflection Analysis ߴɾ࣭ͷΧϯηϦϯάΛൺֱ ߴɾ࣭ͳΧϯηϦϯάΛൺֱ͢Δͱɺൃൃසɺײ ɺ࣌ܥྻͰݟͨײมԽʹେ͖ͳҧ͍͕ݟΒΕΔଞɺ࣭૬खͷҙ ਤΛ܁Γฦ͢Α͏ͳൃͷʹେ͖ͳҧ͍͕ݟΒΕΔ
8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTICFUXFFO)JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT
#Conversation Theme ൃͷςʔϚΛൺֱ 8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTICFUXFFO)JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT είΞ͕ΑΓେ͖͚Ε࣭ͷߴ͍ɺখ͚͞Ε࣭ͷ͍ΧϯηϦϯάͷಛ ߴɾ࣭ͳΧϯηϦϯάͷձςʔϚΛൺֱ͢ΔͱɺNPUJWBUJPO QMBOͳͲͷಛఆͷςʔϚ͕࣭ͷߴ͞ͱ૬͍ؔͯͯ͠ɺ͔ͭΫϥΠΞϯτͱ ίϯαϧλϯτͰςʔϚʹඍົͳҧ͍͕ݟΒΕΔ͜ͱ͕Θ͔Δ
#Feature #Alignment #LSM "MJHONFOUಛͱͯ͠-4. -$.Λ༻ ྔࢺɺଓࢺɺ෭ࢺɺॿಈࢺɺલஔࢺɺףࢺɺਓশ໊ࢺɺ͓Αͼඇਓ শ໊ࢺΛؚΉɺ-*8$ࣙॻʢ5BVTD[JLBOE1FOOFCBLFSɺʣͷ ͭͷݴޠϚʔΧʔશମͰλʔϯϨϕϧͰධՁ 8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTICFUXFFO)JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT
-JOHVJTUJD4UZMF.BUDIJOH -4. -JOHVJTUJD4UZMF$PPSEJOBUJPO -4$ ΧϯηϥʔͱΫϥΠΞϯτͷݴޠతಛͷҰகΛධՁ͢Δ
#Method #Feature Extraction /HSBN୯ޠͱ୯ޠϖΞͷग़ݱස 4FNBOUJD-*8$ 8PSE/FUΛͬͨҙຯతΧςΰϦ༩ .FUBGFBUVSFΧϯηϥʔɾΫϥΠΞϯτͷൃɺ୯ޠɺׂ߹ 4FOUJNFOUϙδωΨχϡʔτϥϧͷఆΛ͠ɺͦΕΒͷೖΕସΘΓ සɺ࣌ܥྻใ "MJHONFOU-4.
-4$ಛ 5PQJDTճؼੳͰಘΒΕͨҼࢠ͝ͱͷείΞ .*5*CFIBWJPSTϦϑϨΫγϣϯϦεχϯάͷׂ߹ 8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTICFUXFFO)JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT ෳͷಛΛఆٛ͠ɺ࣭ͷ༧ଌΛߦͬͨ
#Conclusion ·ͱΊ 8IBU.BLFTB(PPE$PVOTFMPS -FBSOJOHUP%JTUJOHVJTICFUXFFO)JHIRVBMJUZBOE-PXRVBMJUZ$PVOTFMJOH$POWFSTBUJPOT ൃϦϑϨΫγϣϯϦεχϯάɺΞϥΠϝϯτͷҰகͳͲɺײ֮తʹ Θ͔͍ͬͯͨ͜ͱ͚ͩΕͲɺߴ࣭ͳΧϯηϦϯάͷಛΛࣗΒ σʔλΛ࡞ͬͯղੳ͍ͯ͠Δͱ͜Ζ͕Ғ͍ ΧϯηϦϯάͷ࣭Λఔͷਫ਼Ͱ٬؍తʹධՁ͢Δ͜ͱ͕Ͱ ͖ΔΑ͏ʹͳΔ͜ͱͰɺΧϯηϥʔͷڭҭϦΞϧλΠϜϑΟʔυόο ΫʹΑΔࢧԉͳͲʹܨ͛ΒΕΔ͠ɺ·ͨνϟοτϘοτͳͲͰΧϯη
ϦϯάΛ͢Δಓ։͚Δ͔