ࢼߦճ਺ͷ૿΍͠ํ 2021೥౓൛ ཧݚAIP / ౦๺େֶ סݚڀࣨ ਗ਼໺ॢ

ࢼߦճ਺ͱ͸ʁ August 3, 2021 Inui Laboratory 2 ੒ޭͷํఔࣜ [լ୔ 2012] ੒ޭճ਺=ࢼߦճ਺º੒ޭ཰ ੒ޭͱ͸ʁʁ ͱࢥ͏͔΋͠Ε·ͤΜ͕… τοϓձٞ࠾࿥ͱ͔ ίϯϖͰ1Ґͱ͔ͩͱ ࢥ͏͜ͱʹ͠·͠ΐ͏

੒ޭ཰͸ʮӡʯͱʮϝλͳٕज़ʯ • ӡཁૉɿݚڀςʔϚ΍पΓͷ؀ڥ • ͨ·ͨ·ಉ͡ςʔϚͰ࿦จ͕ग़Δͱ͔… • ϝλͳٕज़ɿݴޠԽࠔ೉ͳ҉໧త஌ࣝ • ෼໺/λεΫʹର͢Δ௚ײɺ௕೥ͷ஌ݟɺܦݧɺ೥ͷޭͳͲ • ࠓ೔໌೔Ͱ֫ಘ͢Δͷ͸ࠔ೉ • ےͷྑ͍ݚڀςʔϚΛબͿ͜ͱͰ͋Δఔ౓ίϯτϩʔϧՄ • Y͞ΜͷࢿྉΛࢀরͷ͜ͱ August 3, 2021 Inui Laboratory 3 ੒ޭͷํఔࣜ [լ୔ 2012] ੒ޭճ਺=ࢼߦճ਺º੒ޭ཰ ͜ͷ߲Λίϯτϩʔϧ͢Δͷ͸ࠔ೉ ஫ɿࢲͷओ؍Ͱ͢

ࢼߦճ਺=ख࣋ͪͷ࣌ؒ×Ұճͷࢼߦʹ͔͔Δ࣌ؒ ͳͷͰɺࢼߦճ਺Λ૿΍͍ͨ͠ August 3, 2021 Inui Laboratory 5 ੒ޭͷํఔࣜ [լ୔ 2012] ੒ޭճ਺=ࢼߦճ਺º੒ޭ཰ ͜Ε͸༗ݶɿ ͕͖ࠩͭʹ͍͘ ʢਓੜͳͷͰ…ʣ [ߴ੉ 2016] ͜͜Λ୹͘͢Δඞཁ͕͋Δʂ

ʮࢼߦʹ͔͔Δ࣌ؒʯͷ಺༁ August 3, 2021 Inui Laboratory 6 • ϓϩάϥϜͷ࣮૷࣌ؒ • طଘͷίʔυΛಈ͔ͯ͠ΈΔ • खݩͷΞΠσΞͷ࣮૷ • ϓϩάϥϜͷσόοά • ϓϩάϥϜͷ࣮ߦ࣌ؒ • લॲཧ • Ϟσϧͷֶश • Ϟσϧͷ༧ଌ • ධՁ ͜Εɺӡ͡Όͳٕͯ͘ज़Ͱ͸ʁ

ຊൃදͷ໨త • ࢼߦճ਺Λ૿΍ͨ͢Ίͷ࣮ફతͳʮٕज़ʯΛ঺հ August 3, 2021 Inui Laboratory 7

͓͜ͱΘΓ • ͜ͷൃද͸ࣗ෼ͷֶΜͰ͖ͨTipsΛಠஅͱภݟͰ ·ͱΊͨ΋ͷͰ͢ • ݸਓͷओ؍͕͔ͬ͠Γೖͬͯ·͢ • ʮશ෦Ͱ͖ͳ͍ͱͩΊʯͰ͸ͳ͍ • গͳ͘ͱ΋ࣗ෼͸શ෦͸Ͱ͖ͯͳ͍ • ʮ͓͢͢Ίͷډञ԰঺հʯ͘Β͍ͷؾ࣋ͪͰ ݟͯཉ͍͠ August 3, 2021 Inui Laboratory 8

࣮૷ฤ August 3, 2021 Inui Laboratory 9

ᶃ࣮૷ͷॱ൪ʹؾΛ͚ͭΔ August 3, 2021 Inui Laboratory 10 • Nothing works and everything is too slow à Panic • Simplify model à Go back to basics: bag of vectors + nnet • Make a smaller network and dataset for debugging • Once no bugs: increase model size • Make sure you can overfit to your dataset • Plot your training and dev errors over training iterations • Then regularize with L2 and Dropout • Then do hyperparameter search • Come to OH! ( 3/6/18 Richard Socher Lecture 1, Slide 2 খ͍͞؆୯ͳϞσϧɾػೳͷ࣮૷͔Β࢝ΊΑ͏ খ͍࣍͞ݩɾσʔληοτͰࢼͯ͠ɼόάΛચ͍ग़ͦ͏ όά͕ແ͘ͳͬͨΒϞσϧͷαΠζͱσʔλ਺Λ૿΍ͦ͏ 5SBJOJOHσʔλʹରͯ͠աֶशͰ͖Δ͜ͱΛ֬ೝ͠Α͏ 5SBJOJOHͱ7BMJEͷ-PTTΛಉ͡άϥϑͰ֬ೝ͠Α͏ աֶशΛ֬ೝ͔ͯ͠Βਖ਼ଇԽख๏ʢ-΍%SPQPVUʣΛࢼͦ͏ ࠷ޙʹϋΠύʔύϥϝʔλͷ୳ࡧΛ͠Α͏ https://web.stanford.edu/class/cs224n/lectures/lecture15.pdf ্͔Βॱ൪ʹ΍Δ Լ͔Βࢼͯ͠͏·͍͔͘ͳ͍ à ্ʹ໭ͬͯ΍Γ௚͠ʢख໭Γʣ͕ൃੜ ࣌ؒͷແବ

ྫ͑͹ɿ෼ྨ໰୊ͷ৔߹ • Ϟσϧ: Bag-of-Words • ࣍ݩͷେ͖͞: ~30࣍ݩ • Optimizer: SGD (lr=0.01) or Adam (α=0.0001) • ύϥϝʔλॳظԽ: ϑϨʔϜϫʔΫͷσϑΥϧτ • ਖ਼ଇԽ(dropoutͱ͔): ઈରʹ͠ͳ͍ • σʔλαΠζ:10Πϯελϯε • όοναΠζ:10 • ͭ·Γɺ1όονʹରͯ͠loss͕0ʹͳΔ͔֬ೝ͢Δ • Epoch਺: ~10 • CPU/GPU: ྆ํͰಈ͔͘ʁ August 3, 2021 Inui Laboratory 11

ᶄίʔυΛ͏·͘෼ׂ͢Δ • લॲཧͷྫ • Sentence Split, Tokenization, POS tagging, Lemmatization, Dependency Parsing, NER August 3, 2021 Inui Laboratory 12 ੜσʔλ લॲཧࡁσʔλ Ϟσϧ લॲཧ ܇࿅ ػցֶश͋Δ͋ΔύΠϓϥΠϯ

΍ͬͯ͸͍͚ͳ͍ɿ࠷ڧtrain.py August 3, 2021 Inui Laboratory 13 ੜσʔλ લॲཧࡁσʔλ Ϟσϧ લॲཧ ܇࿅ શ෦train.py Ͱ࣮ݱͰ͖ͦ͏ʁʂ ๻͸͜ΕΛ࠷ڧUSBJOQZͱݺΜͰ͍·͢

͋Γ͕ͪͳ࠷ڧtrain.pyͷҰ෦ • ຖճMeCabͰ෼͔ͪॻ͖ͯ͠ΔͶ… August 3, 2021 Inui Laboratory 14 EFGSFBE@EBUBTFU QBUI EBUBTFU<> XJUIPQFO QBUI BTGJ GPSMJOFJOGJ EBUBTFUBQQFOE .F$BCQBSTF MJOFTUSJQ SFUVSOEBUBTFU

࠷ڧtrain.pyΛආ͚Δ΂͖Nݸͷཧ༝ • ֶशͷͨͼʹಉ͡લॲཧ͕૸Δ • ࣌ؒͷແବ • ܭࢉࢿݯͷແବ • ిؾ୅ͷແବ • ಉ͡ʮલॲཧʯʹͳΔͱ͸ݶΒͳ͍à όάͷԹচ • ϥΠϒϥϦͷόʔδϣϯ΍Φϓγϣϯҧ͍ • ྫ: Stanford CoreNLP • ྫ: MeCabͷࣙॻɿJuman? IPAdic? Unidic? • લॲཧ͕ҧ͏ à ग़དྷ্͕Δσʔλ΋ҧ͏ • ͭ·Γग़དྷ্͕ΔϞσϧ΋ҧ͏ • લॲཧ͸Ұ౓͚ͩʹͯ͠ɺಉ͡σʔλΛ࢖͍ଓ͚Δ΂͖ August 3, 2021 Inui Laboratory 15

ᶅσʔλͷܗࣜʹؾΛ͚ͭΔ • CSV΍TSVɿ΍ΊΑ͏😭 • ಛʹࣗલͰreader/writerΛॻ͘ͷ͸΍ΊΑ͏ • Pythonͷࣙॻͱͷ૬ੑ͕ѱͯ͘໘౗ • JSON Linesɿ͍͍ͧ😄 • ཁ͢Δʹ1ߦ1json • json.dumps / json.loads Ͱ؆୯ʹಡΈॻ͖Ͱ͖Δ August 3, 2021 Inui Laboratory 17

ᶆੵۃతʹassertionΛ࢖͓͏ • ʮ͜͜͸ઈର͜ͷ৚͕݅ຬͨ͞ΕͯΔ͸ͣʂʯ ͱ͍͏ՕॴʹೖΕ͓ͯ͘ • Ξϗͳ৚݅Ͱ͋Δ΄Ͳྑ͍ͱࢥ͏ • ͍͔ͭ݁ہྫ֎Λ౿Ή͜ͱʹͳΔ • ྫ͑͹… • ߦྻԋࢉޙͷshapeͷ֬ೝʢίϝϯτΑΓ΋assertionʣ • ϋΠύϥͷ֬ೝ (AdamͰalpha=0.1ͱ͔Λճආʣ August 3, 2021 Inui Laboratory 18 BTTFSUMFO TSD@EBUB MFO USH@EBUB EncDecͷ࣮૷Ͱ Α͘࢖͏

ᶇargparseΛ࢖͏ • ίʔυͷதʹม਺Λϕλॻ͖͢Δͷ͸΍ΊΔ • ίʔυͷ࠶ར༻ੑ͕མͪΔ • argparseߏจΛຖճॻ͘ͷ͕໘౗ͳΒɺΤσΟλ ͷςϯϓϨʔτ΍εχϖοτʹొ࿥͓ͯ͘͠ • choices/default/required option ͸೺Ѳ͓ͯ͘͠ August 3, 2021 Inui Laboratory 19 γϯϓϧʹ஍ࠈ ݟ௨͕͠ྑ͍

ᶈਓʹݟͤΒΕΔίʔυʹ͓ͯ͘͠ • ʮࣗ෼ͷίʔυΛଞਓ͕ݟΔػձͳ͍͚Ͳʁʯ • ·͋ɺ͔֬ʹݚڀ͸ݽಠͳӦΈ͚ͩͲ… • ͍͍͑ɺ͋Γ·͢ʂ • Ұिؒޙͷࣗ෼͸ଞਓͰ͢ • ࡉ͔͍࢓༷΍ҙਤ͸͙͢ʹ๨Εͯ͠·͏ • ͳͷͰ… • ͨ͘͞ΜίϝϯτΛೖΕ͓ͯ͘ • ҙຯͷ͋Δม਺໊Λ͚͓ͭͯ͘ • ݴޠॲཧ100ຊϊοΫ͸ྑ͍࿅श • ଞਓ͕ݟͯͻͱ໨Ͱ෼͔Δίʔυͩͱྑ͍ • ·ͣ͸ϦʔμϒϧίʔυΛಡΜͰΈΑ͏ August 3, 2021 Inui Laboratory 20

ᶉॲཧঢ়گΛՄࢹԽ͠Α͏ • ࣌ؒͷ͔͔ΔforจͰ͸progressbar/tqdmΛ࢖͏ • ͪΌΜͱॲཧ͕ਐΜͰ͍Δ͜ͱ͕෼͔Δ • ඞཁͳ࣌ؒ΍ॲཧର৅ͷ௕͔͞Β෼͔Δόά΋͋Δ • ʁʁʁʮલॲཧʹ24࣌ؒʁົͩͳ…ʯ • ʢόάʹؾ͕ͭ͘·Ͱʹʣ଴ͭ࣌ؒ΋ݮΔ August 3, 2021 Inui Laboratory 21

ᶊϓϩάϥϛϯά༻ͷϑΥϯτΛ࢖͏ • 1ͱlͷ۠ผɺoͱOͱ0ͷ۠ผɺIͱ|ͷ۠ผ͸ॏཁ • λʔϛφϧͱΤσΟλʹઃఆ͓ͯ͘͠ • Source Han Code JPɺRictyɺHack͋ͨΓ͕༗໊ ͔ͳʁ • ͜͜Β΁Μ͸޷Έͷ໰୊à ޷͖ͳ΍ͭΛ࢖͏ͷͰOK • Α͘Θ͔ΒΜͳΒ: Ricty August 3, 2021 Inui Laboratory 22

ᶋਪ࿦Ͱ͖ΔΑ͏ʹ͓ͯ͘͠ • ਪ࿦Λ͢ΔͨΊͷεΫϦϓτ͸ࠓ͙͢ʹॻ͜͏ • ଞͷ࿦จͱൺֱͰ͖ΔΑ͏ʹ͠Α͏ • ܇࿅͚ͩͰ͖ͯ΋࿦จ͸ॻ͚ͳ͍ʂ • ͜ͷ࡞ۀΛʮτϯωϧͷ։௨ʯͱ๻͸ݺΜͰ͍·͢ • %0/& ։௨ JTCFUUFSUIBO1&3'&$5 ͖Ε͍ͳίʔυ • ͬͯβοΧʔόʔά΋ݴͬͯͨΑʜ • ࠷௿ݶඞཁͳεΫϦϓτͨͪ • อଘͨ͠ύϥϝʔλΛಡΈࠐΜͰ • ະ஌ͷσʔλʹରͯ͠ • ༧ଌ݁ՌΛฦ͠ • είΞʢFH '஋ #-&6 ਖ਼ղ཰ʣ Λܭࢉ͢Δ August 3, 2021 Inui Laboratory 23

ᶋਪ࿦Ͱ͖ΔΑ͏ʹ͓ͯ͘͠ August 3, 2021 Inui Laboratory 24 ੜσʔλ લॲཧࡁσʔλ Ϟσϧ લॲཧ ܇࿅ ᶃ Ϟσϧͷֶशʢ͜͜·ͰͰຬ଍͕ͪ͠ʣ ᶄ ϞσϧͷධՁʢͬͪ͜΋ॏཁʂʣ ςετσʔλ ܇࿅ࡁϞσϧ ೖྗ ༧ଌ ධՁ Compute_acc.py ਖ਼ղ཰ 80% ධՁ εΫϦϓτ

͓·͚ɿܭࢉػ͸ৗʹಈ͔͓ͯ͘͠ • 8͔͔࣌ؒΔ࣮ݧ͕͋ͬͨͱ͢Δ August 3, 2021 Inui Laboratory 25 ͜͜Ͱ࣮ݧ͕ऴΘΔͷΛ଴͍ͬͯΔͷ͸໪ମͳ͍ ேʹ͸લ೔ͷ࣮ݧ͕ऴΘ͍ͬͯΔͱྑ͍

͓·͚ɿܭࢉػ͸ৗʹಈ͔͓ͯ͘͠ • 8͔͔࣌ؒΔ࣮ݧ͕͋ͬͨͱ͢Δ August 3, 2021 Inui Laboratory 26 ͜ͷ࣌ؒΛ࢖࣮ͬͯݧΛճ͓͚ͯ͠Δͱྑ͍ ࣗ෼ͷٳΈ࣌ؒʹܭࢉػΛ͍͡Ίൈ͘Πϝʔδ "Your computer should always be running experiments, since there's no reason that you should be working harder than your computer... On the other hand, just because your computer is running, it doesn't mean that actual progress is being made on the research, if your code/scripts have bugs!"

͓·͚ɿܭࢉػ͸ৗʹಈ͔͓ͯ͘͠ • 8͔͔࣌ؒΔ࣮ݧ͕͋ͬͨͱ͢Δ August 3, 2021 Inui Laboratory 27 ͭ·Γɺ ؼ୐͢Δલʹ࣮ݧΛ౤ೖͰ͖ΔͱҰ൪ྑ͍

ίʔυ؅ཧฤ August 3, 2021 Inui Laboratory 28

αʔόͱϩʔΧϧͷ΍ΓऔΓ: SFTPҰ୒ • αʔόͰશͯͷίʔυΛॻ͘ • 😀ϩʔΧϧ͔ΒίʔυΛҠ͢ख͕ؒແ͍ • 😩αʔόͷࢮ⇛ݚڀɾଔ࿦ͷࢮʢσΟεΫ͸ࢮ͵΋ͷʣ • ϩʔΧϧͰίʔυΛॻ͖ɼखಈͰαʔόʹίϐʔ • 😀ϩʔΧϧͷίʔυ͸৑௕ԽͰ͖ΔʢDropboxͳͲʣ • 😩αʔό΁ͷίϐʔ͕໘౗ʢˍ๨ΕΔʣ • ϩʔΧϧͰίʔυΛॻ͖ɼࣗಈͰαʔόʹίϐʔ • 😀ϑΝΠϧอଘͷ౓ʹཪͰಉظ͕૸Δ • ΤσΟλ໊ + SFTPͰݕࡧͯ͠ΈΑ͏ • PyCharm, Atom, VSCodeͳΒͰ͖Δ͸ͣ August 3, 2021 Inui Laboratory 29 ܁Γฦ͞ΕΔ࢒೦ͳޫܠ ຖճscp΍rsyncͯ͠ͳ͍ʁ

۩ମྫ: ਗ਼໺ͷ৔߹ August 3, 2021 Inui Laboratory 30 ͜Ε΋׬ᘳͰ͸ͳ͍͕ɼ͜͜4೥Ҏ্͸͜ΕͰམͪண͍͍ͯΔ EncDecAttn EncDecAttn deploy/EncDecAttn (SFTP) push pull git git MyPC GitHub (private repo) Dropbox dev ݚڀϓϩδΣΫτ͝ͱʹ͜ͷઃఆΛ͓ͯ͘͠ ׳ΕΕ͹30ඵͰͰ͖Δ ͜͜Ͱ։ൃͱ σόοά

όʔδϣϯ؅ཧɿ΍Δ΂͖ • GitΛ࢖͓͏ • ίϛοτϝοηʔδ͕update΍fixͩΒ͚Ͱ΋େৎ෉ • add, commit, push, pull͕࢖͑Ε͹OK • GitίϚϯυΛ࢖͏ΑΓɼGUIͷΞϓϦΛ࢖͓͏ • Ωʔϫʔυ: SourceTree, Fork, GitGraken • ίϚϯυ͸ෆඞཁʹ೉͍͠ • ڞಉݚڀऀʹݟ͑ΔΑ͏ʹίʔυΛڞ༗͠Α͏ • GitHubΛ༗ޮʹ࢖͓͏ • PrivateϦϙδτϦΛ࡞Ζ͏ • ࠓ͙͢GitHubͷEducationalΞΧ΢ϯτΛ࡞Ζ͏ʢແྉͩΑʂʣ • ະൃදͷݚڀ੒ՌΛ1VCMJDͳܗͰެ։͢Δͷ͸΍ΊΑ͏ • ݚڀࣨͷGitHubάϧʔϓΛੵۃతʹ׆༻͠Α͏ • ݚڀࣨͷਓ͸σϑΥϧτͰશһݟΒΕͯศར August 3, 2021 Inui Laboratory 31 ʮࠓͷίʔυ͕ಈ͔ͳ͘ͳΔͷ͕ݏͰɼมߋΛՃ͑ͮΒ͍ʯ͜ͱ͕ͳ͍Α͏ʹ

Git: CLIΑΓ΋GUI August 3, 2021 Inui Laboratory 32 • ͳΕΔ·Ͱ͸CLIΑΓ΋GUIΛ࢖͓͏ • CLI͸֮͑Δ΂͖͜ͱ͕ଟ͗͢Δ • add, commit, push, fetch, pull, … • GUIΫϥΠΞϯτ͸ࢹ֮తʹΘ͔Γ΍͍͢ Θ͔ΒΜ Θ͔Δʂ

ݚڀࣨͷGitHubάϧʔϓΛ࢖͏ • ࣗ෼ͷݚڀ༻ͷϦϙδτϦ͸cl-tohokuάϧʔϓʹ࡞͓ͬͯ͘ • ݚڀࣨͷਓͳΒ୭Ͱ΋ݟΒΕΔͷͰศརʢট଴ͷख͍ؒΒͣʣ • ஫ɿϓϥΠϕʔτϦϙδτϦʹ͓ͯ͘͜͠ͱ • ະൃදͷݚڀ੒ՌΛެ։৘ใʹͯ͠͸͍͚ͳ͍ August 3, 2021 Inui Laboratory 33

σΟϨΫτϦߏ଄ʹؾΛ഑Ζ͏ ࢲͷ৔߹ workspace src lib util showcase cabocha nlp100 ml euler experimetal search lib target bin data project main test java script src main test java scala scala (JUͰ͸؅ཧ͠ͳ͍ tmp ϓϩδΣΫτ ୯ҐͰ (JU؅ཧ August 3, 2021 Inui Laboratory 34 ݚڀͷ؅ཧ [দྛ 2017] http://www.cl.ecei.tohoku.ac.jp/local/handouts/2017/misc/misc-20171120-y-matsu.pdf Կ͕Կ΍Β… 😩 Θ͔Γ΍͍͢😄 GitͰ؅ཧ͢Δର৅Λ෼͔Γ΍͘͢

੒ޭʢࣦഊʣΛ࠶ݱ͠Α͏ • ࣮ݧϩάΛ࢒ͦ͏ɽ • ྫ͑͹: {࣮ݧͷ೔࣌ɼσΟϨΫτϦ໊ɼίϚϯυ໊ɼ argparseͷத਎ɼPythonͷόʔδϣϯɼϥΠϒϥϦͷ όʔδϣϯɼGitͷίϛοτID} ΛϑΝΠϧʹॻ͖ग़͢ • ϩάΛ࢒͢ͱ͖: printΑΓ΋loggerΛ࢖͓͏ • Pythonඪ४ͷloggerͰ΋ྑ͍͠ɼlogzeroͷΑ͏ͳ ϥΠϒϥϦΛ࢖ͬͯ΋ྑ͍ • ඪ४ग़ྗ/Τϥʔग़ྗΛ࢖͍෼͚Α͏ • teeΛ׆༻͠Α͏ August 3, 2021 Inui Laboratory 35 ྫ: ਗ਼໺ͷ࣮ݧεΫϦϓτ ???ʮState-of-the-Artग़ͨΜ͚ͩͲͳʙʯ

ಓ۩ཱͯฤ ศརͳಓ۩Λ࢖͓͏ August 3, 2021 Inui Laboratory 36

ZSHͷϓϩϯϓτΛ੔උ͢Δ August 3, 2021 Inui Laboratory 38 ΧϨϯτσΟϨΫτϦ ͱ Gitͷϒϥϯν ͱ PythonͷԾ૝؀ڥ Λৗʹग़͓ͯ͘͠ ͱͬͯ΋؆ૉ શͯΛهԱ͢Δྗ͕ඞཁ Α͘Θ͔Βͳ͍ͳΒspaceship-promptΛೖΕ͓ͯ͘

ZSHʹΑΔݡ͍ʁิ׬ • ZSHͷੈքɿجຊ͸λϒ࿈ଧ • λϒҎ֎ΛଧͬͨΒෛ͚ͷੈք؍ • ʮZsh λϒ ิ׬ʯͰݕࡧ͢ΔͱͳΜ͔ग़·͢ August 3, 2021 Inui Laboratory 39

SSHͷconfigΛॻ͘ • https://io-lab.esa.io/posts/1973 Λࢀߟʹઃఆ • ͲͷωοτϫʔΫ͔ΒͰ΋ ssh <αʔό໊> Ͱ઀ ଓͰ͖Δ • αʔόͰjupyter notebookΛಈ͔ͯ͠ खݩͰͭͳ͙͜ͱ΋Մೳ August 3, 2021 Inui Laboratory 40

ΠϯλϥΫςΟϒϑΟϧλ • pecoͱ͔fzfͱ͔ • ઃఆ: άάΖ͏ August 3, 2021 Inui Laboratory 41 ίϚϯυཤྺͷิ׬ σΟϨΫτϦؒͷҠಈ

৘ใऩूฤ August 3, 2021 Inui Laboratory 51

৘ใऩूΛ͠Α͏ • Slack • #misc_python, #misc_tech, #dep_info_sys, #b_sig_nn, #b_sig_dl_club, #_research_skills, • #g_timelineΛͨ·ʹݟΔ • ༗ӹtips͕঺հ͞Ε͍ͯΔ͜ͱ΋ • Twitter • ͜ͷล͸লུ • ݚڀࣨͷաڈͷݚڀTipsͳͲ August 3, 2021 Inui Laboratory 52 TwitterͰݚڀ͕ਐΉʢ͜ͱ͕͋Δʣ

ʮਖ਼͍͠ʯ৘ใऩूΛ͠Α͏ • ΅͘ʮxxxͬͯπʔϧΛαʔόʹೖΕͯΈΑ͏ʯ • 😩 QiitaͷهࣄΛಡΜͰࢼ͢ • ؅ཧऀݖݶ(sudo)Λ࢖͍ͬͯΔ͜ͱ͕ଟ͍ • sudo෇͖ͷίϚϯυ͸αʔόʔͰ͸࢖͑·ͤΜ • هࣄͷ࣭͕ۄੴࠞަ • ެࣜυΩϡϝϯτͷޡ༁هࣄ • ୈҰઢͷݚڀऀ͕ॻ͍ͨղઆهࣄ΋͋Δ͕… • 😀ެࣜυΩϡϝϯτʹैͬͯೖΕΔ • ։ൃऀ(ୡ)͕ॻ͍ͯΔͷͰ৴༻Ͱ͖Δ • ӳޠΛڪΕͳ͍৺͕ॏཁ • ڧ͘Φεεϝɿ4UBDL0WFSGMPX • ӳޠʹͳΜͯෛ͚ͳ͍ʂ August 3, 2021 Inui Laboratory 53 ޡͬͨهࣄͷ্Ͱࢼߦࡨޡ͢Δͷ͸͕࣌ؒ΋͍ͬͨͳ͍

ࢼߦͷճ਺ΛݮΒͦ͏ʢαϘΖ͏ʣ • ࢼߦ͠ͳͯ͘ྑ͍΋ͷ͸αϘΓํ΋ߟ͑Δ • ϋΠύʔύϥϝʔλ͸࿦จ΍ஶऀͷίʔυʹॻ͍ͯ͋Δ • ࿦จΛಡΈղ͘ΑΓ΋ɼ࣮૷ΛಡΜͩ΄͏͕଎͍͜ͱ΋ • ྫ: Attention is All You Needʹର͢ΔɼAnnotated Transformer (http://nlp.seas.harvard.edu/2018/04/03/attention.html) • ࿦จͱ࣮૷͸྆ྠ • μϝݩͰஶऀʹϝʔϧͯ͠ΈΑ͏ • ࣮ݧͷࡉ͔͍ઃఆʢe.g. ϋΠύϥʣΛ஌Γ͍ͨͱ͖ʹ • ʮ࿦จʹ͸ॻ͖͖Εͳ͍ࡉ͔͍ઃఆʯ͸ࢁఔ͋Δ… • લॲཧࡁΈͷϑΝΠϧ͕ཉ͍͠ͱ͖ʹ • ࣗ෼ͷ৔߹: Discourse TreebankͷϑΝΠϧΛ໯ͬͨΓ • (JU)VCͷ*TTVFΑΓ΋ϝʔϧͰ࿈བྷ͠Α͏ • ฦ৴͕དྷͳͯ͘΋ٽ͔ͳ͍ August 3, 2021 Inui Laboratory 54

ݚڀࣨͷ͋Δ͖͔ͨฤ August 3, 2021 Inui Laboratory 55

ੵۃతʹਓʹ૬ஊ͠Α͏ • 10෼೰ΜͰ෼͔Βͳ͔ͬͨΒਓʹ૬ஊ͠Α͏ • େ఍ͷ৔߹ɼ͙͢ʹղܾ͢Δ or ͱ͔͔ͬΓ͕௫ΊΔ • ʮxxxͬͯ࿦จʹؔ࿈͢Δ͔΋…ʯ • ʮੲͦͷݚڀͯͯ͠ɼyyyʹσʔλ͕͋Δ͔Β…ʯ • ʮzzzͬͯϝιου͕͋Δ͔ΒͦΕͰ…ʯ • ఆظMTGͰ࣭໰͢Δͷ͸΋͍ͬͨͳ͍→Slackʹॻ͜͏ • ʢ͜Ε͸ࣗ෼΋΍Γ͕͚ͪͩͲ…ʣ • ਓʹ࿩ͯ͠/Slackʹॻ͍͔ͯΒ͓͔͠͞ʹؾ͕ͭ͘͜ͱ΋͋Δ • ΞώϧਓܗͱσόοΫ͢ΔΈ͍ͨͳจԽ΋͋Δ • ͨͩ͠ɼ૬ஊ͢Δͱ͖͸… • ΤϥʔϝοηʔδશମͱݩͷίʔυΛఴ෇͠Α͏ • 😩: ͳΜ͔͏·͍͖͘·ͤΜͰͨ͠/ಈ͖·ͤΜͰͨ͠ • 😀: train.pyΛ૸Βͤͨͱ͖ͷΤϥʔશจ͕͜ΕͳΜͰ͕͢… • ໰୊Λ࠶ݱ͢Δ࠷খͷϓϩάϥϜɾೖग़ྗΛఴ෇͠Α͏ • ૬ஊ಺༰ͱ૬ஊର৅ ਓ ͷϚοϐϯάΛ࡞Ζ͏ • αʔόͷ͜ͱ͸Y͞ΜɼPythonͷ͜ͱ͸T͞ΜɼͳͲ… August 3, 2021 Inui Laboratory 56

૬ஊɿόουύλʔϯ August 3, 2021 Inui Laboratory 58 Ͷͪ͜ΌΜ ݚڀࣨͷϑϨϯζ ͳΜ͔ಈ͖·ͤΜͰͨ͠ʂ Ͳ͏ͨ͠Β͍͍Ͱ͠ΐ͏͔ʁʁ ʢ͑͐ʜʣ

૬ஊࣄͷࡾ఺ηοτ Τϥʔϝοηʔδશମ • ϓϩάϥϜͷు͖ग़ͨ͠ϩάΛશͯࡌͤΔ • লུ͠ͳ͍ • εΫγϣΑΓ΋ςΩετͰʂʢը૾Ͱ͸άάΕͳ͍ʣ ΤϥʔΛ࠶ݱ͢Δ࠷খͷίʔυɾ࣮ߦྫ • ਺ඦߦͷίʔυ͸ಡΈͨ͘ͳ͍ • ࠷খʹ͢Δͷ͕೉͍͠ͱ͖΋͋Δ͕ɺ࠷௿ݶ͕Μ͹Δ • ࢼ͢ଆͱͯ͠͸ɺίϚϯυҰൃͰ࠶ݱͰ͖Δͱخ͍͠ ࢼͨ͠಺༰ • ʮxxxΛٙͬͯyyyΛ΍ͬͨΜͰ͕͢zzzͰͨ͠ʯ August 3, 2021 Inui Laboratory 59

August 3, 2021 Inui Laboratory 60 ָ͍͠ࢼߦͰ ָ͍͠ݚڀࣨϥΠϑΛʂ

ࢀߟจݙ • ࢼߦճ਺Λ૿΍͍ͨ͠ • ੒ޭͷํఔࣜ (https://www.jstage.jst.go.jp/article/jnlp/19/1/19_1/_pdf) • χϡʔϥϧωοτϫʔΫͷϑϨʔϜϫʔΫͷ࿩ (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2016/seminar/seminar-20161013-takase.pdf) • ࣮૷ͷॱ൪ʹؾΛ͚ͭΑ͏ • Natural Language Processing with Deep Learning (Lecture 15) (https://web.stanford.edu/class/cs224n/lectures/lecture15.pdf) • σΟʔϓϥʔχϯάͷϑϨʔϜϫʔΫΛ࢖ͬͨࣗવݴޠॲཧݚڀͷ࣮૷ͷ஌ݟ (https://www.slideshare.net/HirokiTeranishi1/ss-79363691) • ػցֶशͰٽ͔ͳ͍ͨΊͷίʔυઃܭ (https://www.slideshare.net/takahirokubo7792/ss-65413290) • ػցֶशͰٽ͔ͳ͍ͨΊͷίʔυઃܭ 2018 (https://www.slideshare.net/takahirokubo7792/2018- 97367311) • Writing Code for NLP Research (https://docs.google.com/presentation/d/17NoJY2SnC2UMbVegaRCWA7Oca7UCZ3vHnMqBV4S Uayc/edit#slide=id.p) • ઌʹग़དྷΔ͜ͱ͸ઌʹऴΘΒͤΑ͏ • ݚڀऀྲྀίʔσΟϯάͷۃҙ (http://www.chokkan.org/publication/coding-for-researchers.pdf) • όάͷೖΓʹ͍͍͘ίʔυΛॻ͜͏ • ઓ͏ͨΊͷϓϩάϥϛϯά (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2015/seminar/seminar- 20160119-tianran.pdf) • ݡ͘αʔόͱϩʔΧϧΛߦ͖དྷ͠Α͏ • ݚڀͷ؅ཧ (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2017/misc/misc-20171120-y- matsu.pdf) August 3, 2021 Inui Laboratory 61

ࢀߟจݙ • σΟϨΫτϦߏ଄ʹؾΛ഑Ζ͏ • ݚڀऀྲྀίʔσΟϯάͷۃҙ (http://www.chokkan.org/publication/coding-for-researchers.pdf) • ݚڀͷ؅ཧ (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2017/misc/misc-20171120-y-matsu.pdf) • A Research to Engineering Workflow | Dustin Tran (http://dustintran.com/blog/a-research-to-engineering-workflow) • ศརͳಓ۩Λ࢖͓͏ • EffectiveSSH (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2016/seminar/seminar-20161128-nakayama.html) • ΠϯλϥΫςΟϒϑΟϧλ (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2016/seminar/seminar-20160715- yokoi.pdf) • Python Packages for Research (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2018/misc/misc-20180702- kiyono.pdf) • ࡞ۀ͕ര଎ʹͳΔλʔϛφϧͷγϣʔτΧοτΩʔ·ͱΊ | bacchi.me (https://bacchi.me/linux/terminal-tips/) • ৘ใऩूΛ͠Α͏ • ಡΉ͜ͱͱه࿥͢Δ͜ͱ (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2015/seminar/seminar-20160125- sosuke.k2.pdf) • AHC-Lab M1ษڧձ ࿦จͷಡΈํɾॻ͖ํ (https://www.slideshare.net/ShinagawaSeitaro/ahclab-m1) • ݡ͘ίʔυΛ؅ཧ͠Α͏ • A Research to Engineering Workflow | Dustin Tran (http://dustintran.com/blog/a-research-to-engineering-workflow) • ੒ޭʢࣦഊʣΛ࠶ݱ͠Α͏ • ࣮ݧεΫϦϓτͷجຊ3఺ηοτ (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2015/disco/disco-20160124- sosuke.k.html) • ࣮ݧͷࣄނΛ๷͍͗ͨ (http://www.cl.ecei.tohoku.ac.jp/local/handouts/2018/seminar/seminar-20180712-asano.pdf) August 3, 2021 Inui Laboratory 62