AI研究者になる－学生時代の経験から最先端の研究まで－ / 2023-09-27

"*ݚڀऀʹͳΔʵֶੜ࣌୅ͷ ܦݧ͔Β࠷ઌ୺ͷݚڀ·Ͱʵ ੴؙᠳ໵ େࡕެཱେֶେֶӃ৘ใֶݚڀՊಛ೚ڭत Ѫඤݝཱࠓ࣏౦த౳ڭҭֶߍ

ੴؙᠳ໵ࣗݾ঺հ • Ѫඤݝཱদࢁ੢த౳ڭҭֶߍଔۀ • େࡕ෎ཱେֶ޻ֶ෦ଔۀ ֶ࢜೥ • େࡕ෎ཱେֶେֶӃ޻ֶݚڀՊमྃ
म࢜೥ • ΧΠβʔεϥ΢ςϧϯ޻Պେֶमྃ ത࢜೥ • %',*υΠπਓ޻஌ೳݚڀηϯλʔ ϙευΫ೥ • ΧΠβʔεϥ΢ςϧϯ޻Պେֶ δϡχΞڭत೥ • ೥݄͔Βେࡕެཱେֶ ಛ೚ڭत ߴߍੜͷࠒ X(PPHMF(MBTT X1I%)BU

ओͳ࢓ࣄ͸೥େࡕສതͰలࣔ͢Δ"*ج൫ͷݚڀ։ൃ

ਓͷ໾ʹཱͭ΋ͷɾਓΛ໘ന͕ΒͤΔ΋ͷΛ࡞Δͷ͕޷͖

झຯ͸ཱྀߦɻҰ൪ԕ͍ͱ͜Ζͩͱೆۃʹߦͬͨ͜ͱ΋

ੴؙᠳ໵ւ֎ܦݧ • ֶ෦Ͱ൒೥υΠπɺम࢜Ͱϲ݄ϑϥϯεʹཹֶ • ͦͷޙυΠπͰब৬ͯ͠೥ؒͷݚڀੜ׆ ത࢜ʙڭһ 1TZCFS-BC େֶͷݚڀࣨΛओ࠻ J2--BC
%',*ͷݚڀࣨΛڞಉओ࠻ "MQIBCFO ελʔτΞοϓͷ$30

ΧΠβʔεϥ΢ςϧϯ

ߨԋͷΞ΢τϥΠϯ 5IFPSFUJDBM 4PDJBM 1FSTPOBM ೔ʑਐԽ͢Δ"* "*ݚڀऀͱ͍͏৬ۀ ਓΛݡ͘͢Δ"* "*ݚڀऀʹͳΔ·Ͱ 1SBDUJDBM

ߨԋͷΞ΢τϥΠϯ 5IFPSFUJDBM 1SBDUJDBM 4PDJBM 1FSTPOBM ೔ʑਐԽ͢Δ"* "*ݚڀऀͱ͍͏৬ۀ ਓΛݡ͘͢Δ"* "*ݚڀऀʹͳΔ·Ͱ

ਓ޻஌ೳ "SUJ fi DJBM*OUFMMJHFODF ͱ͸ԿͩΖ͏͔ νϟοτϘοτ͸"*ʁ ࣗಈӡసं͸"*ʁ ࣗಈυΞ͸"* ԿΛ΋ͬͯਓ޻஌ೳͱݺͿ͔͸ݚڀऀͷؒͰ΋ᐆດ
൚༻ɺಛԽɺFUD OBCDPTZTUFN ࠓ೔ͷߨԋͰ͸ʮ஌ੑΛײ͡ΒΕΔٕज़ʯΛਓ޻஌ೳͱ͠·͢ ֶͿɾߟ͑Δɾ఻͑Δೳྗ

༷ʑͳڝٕͰਓؒΛ্ճΔࢥߟೳྗ "MQIB(P ғޟ"*͕ਓؒͷϓϩع࢜ΛഁΔ %FFQ#MVF νΣε"*͕ਓؒͷϓϩع࢜ΛഁΔ

࿭੕Λൃݟ

λϯύΫ࣭ͷߏ଄Λ༧ଌ

පؾͷ਍அɾ࣏ྍิॿ ໢ບը૾͔Β༷ʑͳපؾ ྫύʔΩϯιϯප Λ਍அ ໢ບը૾͔Βͷ൚ԽՄೳͳ ࣬ױݕग़ͷͨΊͷجૅϞσϧ

ੜ੒"* ࣥචɺඳըɺ࡞ۂɺFUD

͜͜਺೥Ͱ"*ݚڀ͕ٸ଎ʹൃలͨ͠ཧ༝͸ʁ :FT /P ೖྗ ग़ྗ ೖྗ ग़ྗ ೖྗ
ग़ྗ ਓؒͷ஌ࣝΛϚχϡΞϧԽ ೴ͷਆܦωοτϫʔΫΛ࠶ݱ ར఺ ܽ఺ ࣮૷ͱཧղ͕ൺֱత؆୯ ϚχϡΞϧʹͳ͍໰୊͸ղ͚ͳ͍ ๲େͳֶशσʔλͱܭࢉࢿݯ͕ඞཁ ະ஌ͷ໰୊΋ֶशͯ͠ղ͚ΔΑ͏ʹ ͜ΕΒ͕ἧͬͨ

࠷৽ͷจॻੜ੒"*͸ԯΛ௒͑ΔύϥϝʔλΛ࣋ͭ Zhao, Wayne Xin, et al. "A survey of
large language models." arXiv preprint arXiv:2303.18223 (2023).

ಛʹ͜͜਺ϲ݄Ͱจॻੜ੒"*͕ൃలͨ͠ཧ༝͸ʁ • େن໛ݴޠϞσϧ -BSHF-BOHVBHF.PEFM ͷొ৔ • ݴޠϞσϧͱ͸ɺςΩετͷଓ͖Λ༧ଌ͢Δ֬཰Ϟσϧ ʮ೔ຊͷट౎͸ʯ ๺ژ
౦ژ ژ౎ ӳࠃͷट౎͸ϩϯυϯͰݕࡧ͢Δͱ݅ ӳࠃͷट౎͸Ͱݕࡧ͢Δͱ ݅ P(ϩϯυϯ|ӳࠃ, ͷ, ट౎, ͸) = 8 21,700 ֬཰͕ΑΓߴ͍୯ޠΛ୳͢ ֬཰͸େྔͷจॻ ίʔύε ͔Βܭࢉ ୯ͳΔ ୯ޠ༧ଌػΛνϡʔχϯά͢Δͱ༷ʑͳ໰୊Λղ͚Δ͜ͱ͕෼͔ͬͨ Ԭ࡚, େن໛ݴޠϞσϧͷڻҟͱڴҖ, AIPγϯϙδ΢Ϝ੒Ռใࠂձ, 2023 https://speakerdeck.com/chokkan/20230327_riken_llm ΑΓൈਮ͠վม

ಛʹ࿩୊ʹͳ͍ͬͯΔ$IBU(15ͱ͸Կʁ • େن໛ݴޠϞσϧ(15 (FOFSBUJWF1SFUSBJOFE5SBOTGPSNFS Λ֦ு • ༩͑ΒΕͨࢦࣔ ϓϩϯϓτ ʹैͬͯԠ౴Λฦ͢Α͏ɺ
  ਓؒͷϑΟʔυόοΫͱڧԽֶश͕૊Έࠐ·Ε͍ͯΔ L Ouyang, J Wu, X Jiang, D Almeida, et. al. 2022. Training Language Models to Follow Instructions with Human Feedback. arXiv:2203.02155. (15ͷճ౴ ϓϩϯϓτ Ϣʔβʔ ཧ૝తͳճ౴ 0QFO"*ࣾͷ Ξϊςʔλʔ ֶश ޻෉ᶃ4VQFSWJTFE fi OFUVOJOH ޻෉ᶄ3FXBSENPEFM 0QFO"*ࣾͷ Ξϊςʔλʔ (15ͷճ౴ (15ͷճ౴ (15ͷճ౴ ճ౴ϥϯΩϯά ֶश (15ͷճ౴

"*ݚڀऀͬͯͲΜͳ࢓ࣄʁ • "*ΛΑΓݡ͘͢ΔͨΊͷΞϧΰϦζϜ ࢉ๏ ΛߟҊͨ͠Γ   "*ٕज़Λ༷ʑͳ෼໺ ҩྍɺڭҭɺ ʹԠ༻ͨ͠Γ͢Δ࢓ࣄ
• ࢓ࣄͷ໨ඪ͸ɺݚڀ੒ՌΛ࿦จ΍੡඼ͱͯ͠ൃද͢Δ͜ͱ • ۈΊઌ͸େֶ ڭҭʹॏ఺ ͔Βاۀ ݚڀʹॏ఺ ·Ͱ༷ʑ • Ͳ͏͢Ε͹ͳΕΔʁത࢜߸ΛऔΔͷ͕Ұൠత ւ֎Ͱ͸ಛʹ • େֶ ֶ࢜೥ େֶӃ म࢜೥ ത࢜೥ Ͱऔಘ͢Δ

ݚڀ׆ಈ • ݚڀςʔϚΛߟ͑Δ • ࿦จΛಡΜͰઌߦݚڀΛௐࠪ͢Δ • ࣮ݧ͢Δ ৘ใֶݚڀͷ৔߹͸γεςϜͷ࣮૷ͱධՁ
• ࿦จΛॻ͘ • ֶձͰݚڀ੒ՌΛ఻͑Δ • Ճ͑ͯɺݚڀࢦಋɾतۀɾֶձӡӦɾֶ಺༻຿ͳͲ

ࠃࡍֶձͷ༷ࢠ

ࠃࡍֶձͷ༷ࢠϓϨθϯςʔγϣϯ

ࠃࡍֶձͷ༷ࢠϙελʔɾσϞൃද

ͳͥ࿦จΛॻ͘ͷ͔ • ۭؒΛ௒͑ͯ஌ࣝΛ఻͑ΔͨΊ • ֶձൃද͚ͩͰ͸ͦͷ৔ʹډ߹Θͤͨਓʹ͔͠఻ΘΒͳ͍ • ࿦จͱͯ͠ൃߦ͞ΕΔͱੈքதͷਓ͕ಡΉ͜ͱ͕Ͱ͖Δ • ࣌ؒΛ௒͑ͯ஌ࣝΛ఻͑ΔͨΊ
• ࿦จ͸ϓϩάϥϜΑΓ௕࣋ͪ͢Δ • ॻ͕ࣜ౷Ұ͞Ε͓ͯΓ৘ใऩू͕ޮ཰తͱ͍͏ϝϦοτ΋

υΠπͰതֶ࢜ੜͱͯ͠ʮಇ͘ʯ • υΠπͷതֶ࢜ੜ͸ϓϩδΣΫτϚωʔδϟ • ए͍େֶੜɾେֶӃੜΛࢦಋ͢Δ ςʔϚܾΊ͔Βʂ • ͦͷܦݧ͕ങΘΕͯमྃޙʹݚڀҎ֎ͷ৬ʹब͘ਓ΋ଟ͍
• څ༩Λ͍ͨͩ͘෼ɺඇৗʹڝ૪త

ւ֎Ͱಇ͖ଓ͚Δ • ݚڀάϧʔϓ಺֎ͷίϛϡχέʔγϣϯ͕࢓ࣄͷத৺ʹ • ༷ʑͳഎܠΛ΋ͭਓͱڠಉ͢Δ௅ઓ • ྫ͑͹ߦಈج४ͷҧ͍ ʮਓʹ໎࿭Λ͔͚ͳ͍ʯͱ  
ʮࣗ෼ͩͬͯਓʹ໎࿭Λ͔͚ΔͿΜਓͷ໎࿭ʹ͸׮༰ʹʯ • ӳޠυΠπޠͰʮڭΘΔʯʮڭ͑Δʯ೉͠͞ͷࠩ

ਓؒͱ"*ͷΑΓྑ͍ڠௐΛߟ͑Δ ੜ·Εͳ͕Βͷ αΠϘʔά ਓͷೳྗͷ௿Լ Digital well-being ೴͸ಡॻػೳΛޙఱత ʹ֫ಘ (೴ͷՄ઼ੑ)

ະདྷ "* ਓ ਓ "* ݡ͞ ࢲͷݚڀςʔϚ   "*ٕज़Ͱਓͷ஌ੑ ֶͼɺߟ͑ɺ఻͑Δྗ
Λ֦ு͢Δ

਎ମతߦಈ e.g., walking, standing, cycling, sleeping ೝ஌తߦಈ e.g., reading, writing,
memorizing, talking ηϯαͱAIͷ૊Έ߹ΘͤͰਓͷߦಈ΍ঢ়ଶΛਪఆ͢Δ ৺ཧతঢ়ଶ  e.g., interest, workload, con fi dence, fatigue

ݚڀ঺հʮֶͿྗʯͷ֦ு ڵຯ ཧղ౓ ೝ஌ෛՙ ΞΠτϥοΩϯά αʔϞάϥϑ ಡΈฦ͠ ඓ෦ද໘Թ౓ ஫ࢹ
S. Ishimaru, et al. Cognitive State Measurement on Learning Materials by Utilizing Eye Tracker and Thermal Camera. Proc. ICDAR HDI 2017, pp. 32–36, 2017. S. Ishimaru, et al. Augmented Learning on Anticipating Textbooks with Eye Tracking. Positive Learning in the Age of Information (PLATO), pp. 387–398, 2018.

ΞΠτϥοΩϯά ςΩετ ਓͷ಺తঢ়ଶ   ڵຯɾཧղ౓   1MFBTFTVNNBSJ[FUIFGPMMPXJOHUFYU   /PUFUIBUUIFSFBEFSJTJOUFSFTUFEJO ϓϩϯϓτ
࢓૊Έ

ࢹઢͰ෼͔Βͳ͍୯ޠΛਪఆͯ͠୯ޠாΛࣗಈͰͭ͘Δ ܭଌ͞Εͨࢹઢ ԁͷେ͖͞஫ࢹ࣌ؒ ೉͍͠ͱײͨ͡୯ޠͷਪఆ ΦϨϯδਖ਼ղ৘ใਫ৭ਪఆ݁Ռ େࣾ et al. "ࢹ఺৘ใΛ༻͍ͨओ؍తߴ೉қ౓୯ޠͷਪఆ". ిࢠ৘ใ௨৴ֶձٕज़ݚڀใࠂ,
vol. 115, no. 517, PRMU2015-189, pp. 149-153, 2016.

ࢹઢͰ֬৴౓Λਪఆͯ͠෮शͷॱ൪Λ࠷దԽ͢Δ ར༻ऀͷֶशσʔλ͕͋Ε͹ͷਫ਼౓ͰਪఆͰ͖Δ ֬৴Λ΋ͨͣʹ౴͑ͯؒҧ͑ͨ໰୊ͷྫ ֬৴Λ΋ͬͯ౴͑ͯؒҧ͑ͨ໰୊ͷྫ S. Ishimaru et al. "Confidence-Aware Learning
Assistant". In arXiv preprint arXiv:2102.07312, 2021.

ݚڀ঺հʮߟ͑Δྗʯͷ֦ு S. Ishimaru, et al. The Wordometer 2.0: Estimating
the Number of Words You Read in Real Life using Commercial EOG Glasses. Proc. UbiComp 2016 Adjunct, pp. 293–296, 2016. S. Ishimaru, et al. Reading Interventions: Tracking Reading State and Designing Interventions. Proc. UbiComp 2016 Adjunct, pp. 1759–1764, 2016. # 3 - Electrodes Horizontal axis: L - R [mV] Vertical axis: B - (L + R)/2 [mV] ؟ిҐ͔Βࢹઢํ޲Λਪఆ 100% 9:41 AM ػցֶशͰʮಡΜͩ୯ޠͷ਺ʯΛਪఆ Ϣʔβʔʹఏࣔ

೥౓ະ౿ࣄۀʮ৺Թܭʯ ηϯαͰه࿥ͨ͠೔ʑͷߦಈϩά͔Β   ৺ͷঢ়ଶΛఆྔԽͯ͠දࣔ͢ΔΞϓϦ

52% 19% 0% 47% 75% 93% 0% 5% 6% Predicted
class low middle high Actual class high middle low BDDVSBDZ S. Ishimaru and K. Kise.“Quantifying the Mental State on the Basis of Physical and Social Activities”. Proc. UbiComp '15 Adjunct, pp. 1217–1220, 2015. ։ൃͷ͖͔͚ͬᶃ ࣗ਎ͷෆௐ͔ΒίϯσΟγϣϯͱηϯασʔλͷؒʹ૬ؔΛൃݟ

։ൃͷ͖͔͚ͬᶄ ମͷෆௐ͸ମԹܭͰ෼͔Δ͕ɺ৺ͷෆௐΛଌΔಓ۩͸ଘࡏ͠ͳ͍

։ൃͷ͖͔͚ͬᶅ ৺ͷ݈߁ঢ়ଶΛश׳తʹख࡞ۀͰه࿥͢Δ͜ͱ͸ࠔ೉

৺Թܭ͸೔ʑͷߦಈྔͷมԽ͔Β৺ͷঢ়ଶΛਪఆ͢Δ ೖྗ ग़ྗ ؾ෼ ػ   ց   ֶ
  श ਎ମߦಈྔ ೝ஌ߦಈྔ ࣾձߦಈྔ ਭ຾ ಛ௃நग़ ׆ྗ iPhone, AppleWatch Fitbit JINS MEME Twitter, Facebook RescueTime ਪఆ

։ൃ੒Ռ म࿦ൃදɾະ౿੒Ռใࠂձલͷ༷ࢠ म࿦ൃද ৺Թܭͳ͠ ະ౿੒Ռใࠂձ ৺Թܭ͋Γ ৺ԹΛνΣοΫ͠ͳ͕Βੜ׆͢Δ͜ͱͰຊ൪·Ͱߴ೤Λ๷͍ͩ

*1"/&847PMIUUQTXBSQOEMHPKQJOGPOEMKQQJEXXXJQBHPKQ fi MFTQEG

ݚڀ঺հʮ఻͑Δྗʯͷ֦ு ൃ࿩ɾᰐ͖ɾস͍Λߴਫ਼౓Ͱਪఆ   'TDPSF Τϯήʔδ౓߹͍ ൃ࿩ɾॻهɾ಺৬ Λߴਫ਼౓Ͱਪఆ
  ϢʔβʔඇґଘͷֶशͰ"DD 'TDPSF Chen, et al. Quantitative Evaluation System for Online Meetings Based on Multimodal Microbehavior Analysis. Sensors and Materials 34 (8), pp. 3017–3027, 2022. Watanabe, et al. EnGauge: Engagement Gauge of Meeting Participants Estimated by Facial Expression and Deep Neural Network IEEE Access, 2023 (Ealy Access).

%JTDVTTJPO+PDLFZ %+ #(.ͳ͠ #(.ݻఆ #(.ಈత ఏҊγεςϜ ൃ࿩ྔʹԠ֤ͯࣗ͡ͷ1$ͷ#(.͕มΘΔ ΦϯϥΠϯϛʔςΟϯάγεςϜΛ։ൃ ಛఆͷࢀՃऀ͕஻Γ͗͢ΔͷΛ཈ࢭ
Ͱ͖Δ͜ͱΛ࣮ݧͰ໌Β͔ʹͨ͠ H. Suzawa et al.Supporting Smooth Interruption in a Video Conference by Dynamically Changing Background Music Depending on the Amount of Utterance. UbiComp '22 Adjunct

෺৺͕͍ͭͨͱ͖͔ΒʮཧՊʯ͕޷͖ͩͬͨ • খֶߍʙߴߍͷཧՊͷ࣮ݧ • ੜ෺ֶΦϦϯϐοΫ • ͓΋͠ΖՊֶίϯςετ • αΠΤϯεΩϟϯϓ
• େֶ๚໰ɾΦʔϓϯΩϟϯύε

໷ߦόεʹ৐ͬͯ౦େͷΦʔϓϯΩϟϯύεʹߦͬͨ

೔ຊ୅ද·Ͱ͋ͱาͩͬͨͷʹ ߴߍͷ)1ʹܝࡌͨ͠ײ૝ΑΓൈਮ ੜ෺ֶΦϦϯϐοΫ̎࣍༧બʢੜ෺νϟϨϯδຊબʣʹࢀՃͰ͖ͯ   Ұ൪ྑ͔ͬͨͷ͸ɺಉ͡ࢤΛ࣋ͬͨ༑ୡ͕શࠃʹͨ͘͞ΜͰ͖ͨ͜ͱͰ͢ɻ   શࠃ͔Βਓͷੜ෺େ޷͖ਓ͕ؒू·͍ͬͯͨͷͰɺΈΜͳͦΕͧΕ   ཱ೿ͳເΛ͍࣋ͬͯ·ͨ͠ɻࢼݧ࠷ऴ೔ʹ͸͓ޓ͍ͷເͱ͔໨ඪΛே·Ͱ  
ޠΓ໌͔͢͜ͱ͕Ͱ͖ɺͱͯ΋͍͍ܹࢗΛड͚·ͨ͠ʂ

ߴߍͰͷߨԋͷޙϝʔϧΛૹ࣭ͬͯ໰ɾେֶ๚໰

ੜ෺ͱ޻ֶͷ༥߹ݚڀΛ͢ΔͨΊʹେࡕେֶجૅ޻ֶ෦Λडݧ

डݧษڧ

डݧʹࣦഊୈࢤ๬ͷେࡕ෎ཱେֶ΁

େֶࡇͷ࣮ߦҕһͱͯ͠ϙελʔ΍ϗʔϜϖʔδ੍࡞Λ୲౰

νίΫΠΠϫέϩϘ SBDPMUB /'$+VODUJPO LPUPEBNB 3FTU$BTU 4XJQFS ͳͲ झຯͰ࡞ͬͨΞϓϦέʔγϣϯ ΞϧόΠτɾΠϯλʔϯγοϓ େֶ೥य़ɿιϑτ΢ΣΞ։ൃ͕झຯɾΞϧόΠτʹ

༑ୡͷձࣾΛख఻ͬͨΓΠϯλʔϯγοϓͰ஥͕ؒͰ͖ͨΓ

։ൃͨ͠ΞϓϦαʔϏε 41"..64&6. ໘ന͓͔͍͠໎࿭ϝʔϧΛ ౤ߘ͢ΔϥϯΩϯάαΠτ 4XJQFS εϫΠϓૢ࡞Ͱ௚ײతʹ ૢ࡞Ͱ͖Δ5P%PΞϓϦ 1BMFUUB ϓϩάϥϚͰ΋͍͍ײ͡ͷ৭Λ
͙͢ʹબ΂ΔΧϥʔύϨοτ

։ൃͨ͠ΞϓϦαʔϏε νίΫΠΠϫέϩϘ ໿ଋͷ࣌ؒʹ஗Εͨͱ͖ͷ ݴ͍༁Λݕࡧ͢ΔΞϓϦ 3FTU$BTU ࣍ʹτΠϨʹߦͩ͘Ζ͏ ࣌ؒΛ༧ใ͢ΔΞϓϦ ,PUPEBNB ৸๥Λޙչͨ͠πΠʔτΛूΊͯ
ਂ໷ʹͻͨ͢Β౤ߘ͢Δ໎࿭#PU

։ൃͨ͠ΞϓϦαʔϏε νίΫΠΠϫέϩϘ ໿ଋͷ࣌ؒʹ஗Εͨͱ͖ͷ ݴ͍༁Λݕࡧ͢ΔΞϓϦ 3FTU$BTU ࣍ʹτΠϨʹߦͩ͘Ζ͏ ࣌ؒΛ༧ใ͢ΔΞϓϦ ,PUPEBNB ৸๥Λޙչͨ͠πΠʔτΛूΊͯ
ਂ໷ʹͻͨ͢Β౤ߘ͢Δ໎࿭#PU ਓͷߦಈ΍ੈͷதͷৗࣝΛม͑Δ΋ͷ͕޷͖   ୭΋͕ϚΠφεͩͱࢥ͍ͬͯΔ΋ͷΛϓϥεʹͰ͖ͳ͍͔

େֶ೥ळɿݚڀࣨ഑ଐσβΠϯɾϓϩάϥϛϯά͕໾ཱͭ

࡞ͬͨ΋ͷΛݟͤΔ͜ͱͰཹֶͳͲͨ͘͞ΜͷνϟϯεΛ͍͍ͨͩͨ

υΠπɾϑϥϯεཹֶͷܦݧΛ௨ͯ͡υΠπ΁ͷब৬ΛܾΊͨ

ݚڀऀʹͳΔ·Ͱͷֶͼ • ࣗ෼͕޷͖ͳ͜ͱͷ࣠Λ࣋ͭ • ΍Γ͍ͨ͜ͱ͸ͲΜͲΜมΘͬͯ΋େৎ෉ • େֶ͸໘ന͍ਓ΍෺ʹग़ձ͑Δ৔ॴ • ͋Ε͜Ε΍͍ͬͯΔͱޙ͔Βܨ͕Δ͜ͱ΋͋Δ
• ϐϯνΛνϟϯεʹม׵͢ΔྗʹͳΔ

ࠓ͔Β࢝ΊΒΕΔ͓͢͢Ί͸ʮະ౿δϡχΞʯ • ಠ૑తΞΠσΞͱ୎ӽٕͨ͠ज़Λ࣋ͭ   খதߴੜͷ׆ಈΛࢧԉ͢ΔϓϩάϥϜ • ୈҰઢͰ׆༂͢Δઌഐ ະ౿ग़਎ऀ ͔ΒΞυόΠε
• ։ൃࢿۚ ্ݶສԁ ͷԉॿ • εʔύʔΫϦΤʔλೝఆ੍

·ͱΊ

"*ݚڀऀʹͳΔͨΊʹ໾ཱͬͨ͜ͱ͸ʁ தֶʙߴߍ • தֶ೥ੜͷͱ͖ʹେֶͷΦʔϓϯΩϟϯύεʹߦͬͨ • डݧ͸ઌͰ΋ʮେֶʯʮݚڀʯͷΠϝʔδΛ͔ͭΊͨ • ՝֎׆ಈʹࢀՃͨ͠ αΠΤϯεΩϟϯϓɺੜ෺ֶΦϦϯϐοΫFUD
• ڵຯ΍ਐ࿏ͷબ୒ࢶ͕૿͑ͨ۩ମԽͨ͠ • ྭ·͠ڝ͍߹͑Δ༑ਓͱஂମઓͷडݧษڧΛͨ͠ • ୈҰࢤ๬ߍʹ͸ߦ͚ͳ͔͕ͬͨɺʮసΜͰ΋ͨͩͰ͸ى͖ ͳ͍ʯੑ֨΍ݚڀʹඞཁͳجૅӳޠྗͳͲ͕਎ʹ͍ͭͨ ·ͱΊ

খ͞ͳ໨ඪΛୡ੒͠ଓ͚Δ • ʮޙչ͠ͳ͍બ୒ʯͰ͸ͳ͘ʮબ୒Λޙչʹ͠ͳ͍ߦಈʯ • ܭըੑΑΓॊೈੑɻ΍Γ͍ͨ͜ͱ͸ͲΜͲΜมΘ͍͍ͬͯ ·ͱΊ ΢ΥʔλʔϑΥʔϧ։ൃ ΞδϟΠϧ։ൃ

"*ݚڀऀʹͳΔʵֶੜ࣌୅ͷ ܦݧ͔Β࠷ઌ୺ͷݚڀ·Ͱʵ ੴؙᠳ໵ େࡕެཱେֶେֶӃ৘ใֶݚڀՊಛ೚ڭत Ѫඤݝཱࠓ࣏౦த౳ڭҭֶߍ ࣭໰ɾײ૝͓଴͓ͪͯ͠Γ·͢ʂ ࠲ஊձ΍!TIPZBJTIJNBSV!PNVBDKQ΁

AI研究者になる－学生時代の経験から最先端の研究まで－ / 2023-09-27

AI研究者になる－学生時代の経験から最先端の研究まで－ / 2023-09-27

More Decks by Shoya Ishimaru

Featured

Transcript