
ãæžãããã¯ããåµãããã
çè«ãåŠã¶ã ãã§ã¯ãçŸå Žã®èª²é¡ã¯è§£æ±ºã§ããŸãããç°å¢æ§ç¯ããé«åºŠãªã¢ãã«å®è£
ãŸã§ã100ã®ã¹ããããéããŠãèªåã®æã§åãå»çAIããäœãäžãããå®è·µçããã°ã©ãã³ã°å€§å
šã§ãã
[Series C] Clinical AI Coding 100 : ã³ãŒã¹ã®å šäœå
第0éšïŒAIãšå»çãåéºã®å§ãŸã

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C0 | 第0åïŒAIãæãå»çã®æªæ¥å°å³ãåéºã¯ããããå§ãŸãïŒ | 人工ç¥èœ(AI), æ©æ¢°åŠç¿, 深局åŠç¿, 第4æ¬¡ç£æ¥é©åœ, Evidence-Based Medicine (EBM), ããžã¿ã«ãã©ã³ã¹ãã©ãŒã¡ãŒã·ã§ã³(DX) |
| C1 | 第1åïŒAIã®ãéªšæ Œããåµããæ°åŠãšããåã®èšèšå³ïœå»çè ã®ããã®5倧ããŒãå ¥é | ç·åœ¢ä»£æ°, 埮å, 確çã»çµ±èš, æé©åçè«, æ å ±çè« |
| C2 | 第2åïŒããŒã¿ãµã€ãšã³ã¹ã®ç¬¬äžæ©ïŒçµ±èšåŠãšç«åŠã§å»çã®ããªãïŒããè§£ãæãã | çµ±èšåŠ, ç«åŠ, å ææšè«, ãã€ã¢ã¹, ç ç©¶ãã¶ã€ã³ |
| C3 | 第3åïŒã¢ã€ãã¢ãã圢ãã«ããéæ³ãããã°ã©ãã³ã° | Python, ã©ã€ãã©ãª, éçºç°å¢, åçŸæ§, èšç®è«çæè |
| C4 | 第4åïŒAIã®ãç®ããè²ãŠããç»åèªèã®ããã³ãã£ã¢ãžããããïŒ | å»çšç»å(DICOM), CNN, ã»ã°ã¡ã³ããŒã·ã§ã³, ç»åèšºææ¯æŽ |
| C5 | 第5åïŒAIã®ãè³ãã§èŽããèšèãšæéã®ã¡ãã㣠| èªç¶èšèªåŠç(NLP), é»åã«ã«ã, çäœä¿¡å·, æç³»åè§£æ |
| C6 | 第6åïŒAIããåµé ããããçæã¢ãã«ãšããåã®ç»æ | çæAI, å€§èŠæš¡èšèªã¢ãã«(LLM), æ¡æ£ã¢ãã«, ããŒã¿æ¡åŒµ |
| C7 | 第7åïŒã詊è¡é¯èª€ãããåŠã¶AIæè¡ã匷ååŠç¿ã«ããæåŒ·ã®æææ±ºå®è¡ | 匷ååŠç¿, 鿬¡æææ±ºå®, åå¥åæ²»ç, åçæ²»çèšç» |
| C8 | 第8åïŒç 究宀ããèšåºãžãAIã瀟äŒã«å±ããéã®ã | MLOps, ãœãããŠã§ã¢å·¥åŠ, èŠå¶å¯Ÿå¿(è¬æ©æ³), ã¢ããªã±ãŒã·ã§ã³éçº |
| C9 | 第9åïŒAIæä»£ã®å»çå«çãå¿ã«å»ãã¹ãã³ã³ãã¹ | AIå«ç, ELSI, 説æå¯èœæ§(XAI), å ¬å¹³æ§, ããŒã¿ãã©ã€ãã·ãŒ, å»çè ãšããŠã®è²¬ä»» |
C0-9
- 第0åïŒC0ïŒAIãæãå»çã®æªæ¥å°å³ãåéºã¯ããããå§ãŸãïŒ
â 人工ç¥èœ(AI), æ©æ¢°åŠç¿, 深局åŠç¿, 第4æ¬¡ç£æ¥é©åœ, Evidence-Based Medicine (EBM), ããžã¿ã«ãã©ã³ã¹ãã©ãŒã¡ãŒã·ã§ã³(DX) - 第1åïŒC1ïŒAIã®ãéªšæ Œããåµããæ°åŠãšããåã®èšèšå³ïœå»çè
ã®ããã®5倧ããŒãå
¥é
â ç·åœ¢ä»£æ°, 埮å, 確çã»çµ±èš, æé©åçè«, æ å ±çè« - 第2åïŒC2ïŒããŒã¿ãµã€ãšã³ã¹ã®ç¬¬äžæ©ïŒçµ±èšåŠãšç«åŠã§å»çã®ããªãïŒããè§£ãæãã
â çµ±èšåŠ, ç«åŠ, å ææšè«, ãã€ã¢ã¹, ç ç©¶ãã¶ã€ã³ - 第3åïŒC3ïŒã¢ã€ãã¢ãã圢ãã«ããéæ³ãããã°ã©ãã³ã°
â Python, ã©ã€ãã©ãª, éçºç°å¢, åçŸæ§, èšç®è«çæè - 第4åïŒC4ïŒAIã®ãç®ããè²ãŠããç»åèªèã®ããã³ãã£ã¢ãžããããïŒ
â å»çšç»å(DICOM), CNN, ã»ã°ã¡ã³ããŒã·ã§ã³, ç»åèšºææ¯æŽ - 第5åïŒC5ïŒAIã®ãè³ãã§èŽããèšèãšæéã®ã¡ããã£
â èªç¶èšèªåŠç(NLP), é»åã«ã«ã, çäœä¿¡å·, æç³»åè§£æ - 第6åïŒC6ïŒAIããåµé ããããçæã¢ãã«ãšããåã®ç»æ
â çæAI, å€§èŠæš¡èšèªã¢ãã«(LLM), æ¡æ£ã¢ãã«, ããŒã¿æ¡åŒµ - 第7åïŒC7ïŒã詊è¡é¯èª€ãããåŠã¶AIæè¡ã匷ååŠç¿ã«ããæåŒ·ã®æææ±ºå®è¡
â 匷ååŠç¿, 鿬¡æææ±ºå®, åå¥åæ²»ç, åçæ²»çèšç» - 第8åïŒC8ïŒç 究宀ããèšåºãžãAIã瀟äŒã«å±ããéã®ã
â MLOps, ãœãããŠã§ã¢å·¥åŠ, èŠå¶å¯Ÿå¿(è¬æ©æ³), ã¢ããªã±ãŒã·ã§ã³éçº - 第9åïŒC9ïŒAIæä»£ã®å»çå«çãå¿ã«å»ãã¹ãã³ã³ãã¹
â AIå«ç, ELSI, 説æå¯èœæ§(XAI), å ¬å¹³æ§, ããŒã¿ãã©ã€ãã·ãŒ, å»çè ãšããŠã®è²¬ä»»
第IéšïŒAIã®æèåè·¯ãèŠããéæ³ã®æ°åŠãšã³ãžã³

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C10 | 第10åïŒAIã®å ¬çšèªããã¯ãã«ãšè¡åã§ããŒã¿ãæããïŒ | ç·åœ¢ä»£æ°, ã¹ã«ã©ãŒ, ãã¯ãã«, è¡å, ãã³ãœã«, ããŒã¿è¡šçŸ, 空é |
| C11 | 第11åïŒããŒã¿ã®æ¬è³ªãèŠæãã倿ãã®æ°åŠãâ è¡åæŒç®ããäž»æååæ(PCA)ãŸã§ | è¡åæŒç®, éè¡å, è¡ååŒ, åºæå€, åºæãã¯ãã«, äž»æååæ(PCA), ç¹ç°å€åè§£ |
| C12 | 第12åïŒAIã¯ããããŠåŠã¶ïŒåŸ®åã®âã³ã³ãã¹âã§åŠã¶ãè³¢ããªãä»çµã¿ | 埮å, åŸé , å埮å, é£éåŸ, åŸé éäžæ³, æå°å€æ¢çŽ¢ |
| C13 | 第13åïŒããã¶ãããç§åŠããã確çã»çµ±èšã®ããã³ | 確ç倿°, 確çååž, æåŸ å€, 忣, æ¡ä»¶ä»ã確ç, ãã€ãºã®å®ç |
| C14 | 第14åïŒAIã®ãèªä¿¡ãã®æºæ³ãå°€åºŠãšæšå®ã®ã»ãªãªãŒ | å°€åºŠé¢æ°, æå°€æšå®æ³, ãã€ãºæšå®, äºåååž, äºåŸååž |
| C15 | 第15åïŒAIãããã¹ããªçãããå°ãåºãçŸ éç€ããæé©åããžã®æ¢æ± | æé©ååé¡, æå€±é¢æ°, åŸé éäžæ³ã®çºå±, åžæé©å, ã©ã°ã©ã³ãžã¥ã®æªå®ä¹æ°æ³ |
| C16 | 第16åïŒæ å ±ã®ã䟡å€ããæž¬ããæ å ±çè«ã®äžæè°ãªäžç | æ å ±é, ãšã³ããããŒ, KLãã€ããŒãžã§ã³ã¹, 亀差ãšã³ããããŒ, çžäºæ å ±é |
| C17 | 第17åïŒAIã¯ãªããå¿çšåé¡ããè§£ããã®ãïŒåŠç¿èœåã®ç§å¯ãè§£ãæããçµ±èšçåŠç¿çè« | æ±åã»éåŠç¿, ãã€ã¢ã¹-ããªã¢ã³ã¹, VC次å , PACåŠç¿, 亀差æ€èšŒ, æ£åå, ã¢ã³ãµã³ãã«åŠç¿ |
| C18 | 第18åïŒãã€ãªããããæ°åŠã§è§£ãæããã倿§äœãã°ã©ãçè«ããããŠGNN | 倿§äœåŠç¿, éç·åœ¢æ¬¡å åæž, t-SNE, UMAP, ã°ã©ãçè«, ãããã¯ãŒã¯åæ, ããŒã/ãšããž, äžå¿æ§åæ, ã³ãã¥ããã£æ€åº, ç¥èã°ã©ã, ã°ã©ããã¥ãŒã©ã«ãããã¯ãŒã¯(GNN) |
| C19 | 第19åïŒAIããæ¬¡ã®äžæããæ±ºããã匷ååŠç¿ã®æ°åŠ | ãã«ã³ã決å®éçš(MDP), ç¶æ ã»è¡åã»å ±é ¬, æ¹ç(Policy), 䟡å€é¢æ°(Value Function), è¡å䟡å€é¢æ°(Q-function), ãã«ãã³æ¹çšåŒ, 䟡å€ååŸ©ã»æ¹çå埩, æéå·®(TD)åŠç¿, QåŠç¿, SARSA, æ¹çåŸé æ³(Policy Gradient), æ¢çŽ¢ãšæŽ»çšã®ãã¬ãŒããªã |
C10-19
- 第10åïŒC10ïŒAIã®å
¬çšèªããã¯ãã«ãšè¡åã§ããŒã¿ãæããïŒ
â ç·åœ¢ä»£æ°, ã¹ã«ã©ãŒ, ãã¯ãã«, è¡å, ãã³ãœã«, ããŒã¿è¡šçŸ, 空é - 第11åïŒC11ïŒããŒã¿ã®æ¬è³ªãèŠæãã倿ãã®æ°åŠãâ è¡åæŒç®ããäž»æååæ(PCA)ãŸã§
â è¡åæŒç®, éè¡å, è¡ååŒ, åºæå€, åºæãã¯ãã«, äž»æååæ(PCA), ç¹ç°å€åè§£ - 第12åïŒC12ïŒAIã¯ããããŠåŠã¶ïŒåŸ®åã®âã³ã³ãã¹âã§åŠã¶ãè³¢ããªãä»çµã¿
â 埮å, åŸé , å埮å, é£éåŸ, åŸé éäžæ³, æå°å€æ¢çŽ¢ - 第13åïŒC13ïŒããã¶ãããç§åŠããã確çã»çµ±èšã®ããã³
â 確ç倿°, 確çååž, æåŸ å€, 忣, æ¡ä»¶ä»ã確ç, ãã€ãºã®å®ç - 第14åïŒC14ïŒAIã®ãèªä¿¡ãã®æºæ³ãå°€åºŠãšæšå®ã®ã»ãªãªãŒ
â å°€åºŠé¢æ°, æå°€æšå®æ³, ãã€ãºæšå®, äºåååž, äºåŸååž - 第15åïŒC15ïŒAIãããã¹ããªçãããå°ãåºãçŸ
éç€ããæé©åããžã®æ¢æ±
â æé©ååé¡, æå€±é¢æ°, åŸé éäžæ³ã®çºå±, åžæé©å, ã©ã°ã©ã³ãžã¥ã®æªå®ä¹æ°æ³ - 第16åïŒC16ïŒæ
å ±ã®ã䟡å€ããæž¬ããæ
å ±çè«ã®äžæè°ãªäžç
â æ å ±é, ãšã³ããããŒ, KLãã€ããŒãžã§ã³ã¹, 亀差ãšã³ããããŒ, çžäºæ å ±é - 第17åïŒC17ïŒAIã¯ãªããå¿çšåé¡ããè§£ããã®ãïŒåŠç¿èœåã®ç§å¯ãè§£ãæããçµ±èšçåŠç¿çè«
â æ±åã»éåŠç¿, ãã€ã¢ã¹-ããªã¢ã³ã¹, VC次å , PACåŠç¿, 亀差æ€èšŒ, æ£åå, ã¢ã³ãµã³ãã«åŠç¿ - 第18åïŒC18ïŒãã€ãªããããæ°åŠã§è§£ãæããã倿§äœãã°ã©ãçè«ããããŠGNN
â 倿§äœåŠç¿, éç·åœ¢æ¬¡å åæž, t-SNE, UMAP, ã°ã©ãçè«, ãããã¯ãŒã¯åæ, ããŒã/ãšããž, äžå¿æ§åæ, ã³ãã¥ããã£æ€åº, ç¥èã°ã©ã, ã°ã©ããã¥ãŒã©ã«ãããã¯ãŒã¯(GNN) - 第19åïŒC19ïŒAIããæ¬¡ã®äžæããæ±ºããã匷ååŠç¿ã®æ°åŠ
â ãã«ã³ã決å®éçš(MDP), ç¶æ ã»è¡åã»å ±é ¬, æ¹ç(Policy), 䟡å€é¢æ°(Value Function), è¡å䟡å€é¢æ°(Q-function), ãã«ãã³æ¹çšåŒ, 䟡å€ååŸ©ã»æ¹çå埩, æéå·®(TD)åŠç¿, QåŠç¿, SARSA, æ¹çåŸé æ³(Policy Gradient), æ¢çŽ¢ãšæŽ»çšã®ãã¬ãŒããªã
第IIéšïŒçµ±èšåŠã§è¬ãè§£ãïŒããŒã¿æ¢åµå ¥éïŒããŒã¿ããçå®ãå°ãåºãç§åŠ

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C20 | 第20åïŒããŒã¿ãšéåŒã«èªãåãæè¡ãçµ±èšçæèãšå¯èŠåã®ç¬¬äžæ© | èšè¿°çµ±èš (代衚å€, æ£åžåºŠ), æšæž¬çµ±èš (æ¯éå£ãšæšæ¬), ããŒã¿å¯èŠå (EDA, ãã¹ãã°ã©ã , ç®±ã²ãå³), 確çã®åºç€ (確ç倿°, æ¡ä»¶ä»ã確ç, ãã€ãºã®å®ç), 確çååž (æ£èŠååž, äºé ååž, ãã¢ãœã³ååž), 倧æ°ã®æ³å, äžå¿æ¥µéå®ç |
| C21 | 第21åïŒãå¶ç¶ããå¿ ç¶ãïŒããèŠæ¥µããç§åŠ â 仮説æ€å®ãšä¿¡é Œåºéã䜿ãããªã | 仮説æ€å® (åž°ç¡ä»®èª¬, 察ç«ä»®èª¬, på€, æææ°Žæº), 第1çš®ã»ç¬¬2çš®ã®é誀, æ€åºå, ä¿¡é Œåºé, tæ€å® (察å¿ãã/ãªã), ã«ã€äºä¹æ€å®, 忣åæ(ANOVA), å€éæ¯èŒ |
| C22 | 第22åïŒããªãïŒãã«çããç§åŠãå ææšè«ãžã®æåŸ ç¶ | çžé¢ãšå æ, 亀絡, 第3ã®å€æ°, æ§é çå æã¢ãã«(SCM), æåéå·¡åã°ã©ã(DAG), dåé¢, ããã¯ãã¢åºæº, åå®ä»®æ³, ããã³ã·ã£ã«ã¢ãŠãã«ã , å æå¹æ (ATE, ATT), èå¥å¯èœæ§ã®3å€§ä»®å® (亀æå¯èœæ§, äžè²«æ§, ããžãã£ããã£) |
| C23 | 第23åïŒæåŒ·ã®ãšããã³ã¹ãåµããèšåºç«åŠãšç ç©¶ãã¶ã€ã³ â ããŒã¿ãããçå®ããèŠæãèªæµ·è¡ | ç«åŠææš (眹æ£ç, æç ç, ãªã¹ã¯æ¯, ãªããºæ¯, å¯äžå±éº), ãã€ã¢ã¹ (éžæãã€ã¢ã¹, æ å ±ãã€ã¢ã¹), ç ç©¶ãã¶ã€ã³ (暪æç ç©¶, çäŸå¯Ÿç §ç ç©¶, ã³ããŒãç ç©¶, ã©ã³ãã 忝èŒè©Šéš/RCT), 蚺æç²ŸåºŠææš (æåºŠ, ç¹ç°åºŠ, ROCæ²ç·, AUC) |
| C24 | 第24åïŒæªæ¥ãäºæž¬ããçµ±èšã¢ããªã³ã°â ïŒååž°åæãã¹ã¿ãŒ | ç·åœ¢éååž°åæ (æå°äºä¹æ³), ããžã¹ãã£ãã¯ååž°åæ (æå°€æ³), ãã¢ãœã³ååž°, ã¢ãã«èšºæ (æ®å·®åæ, å€éå ±ç·æ§/VIF), 倿°éžæ (ã¹ãããã¯ã€ãº, LASSO), ã¢ãã«è©äŸ¡ (決å®ä¿æ°, AIC, AUC) |
| C25 | 第25åïŒæªæ¥ãäºæž¬ããçµ±èšã¢ããªã³ã°â¡ïŒçåæéåæ | æã¡åã (å³åŽ/å·ŠåŽ), çå颿°, ãã¶ãŒã颿°, Kaplan-Meieræ³, ãã°ã©ã³ã¯æ€å®, Coxæ¯äŸãã¶ãŒãã¢ãã«, æ¯äŸãã¶ãŒãæ§ã®ä»®å®, æéäŸåæ§å ±å€é |
| C26 | 第26åïŒãèŠããªãæµã亀絡ãè¯éºã«ãã°ããã¯ãã㯠| å±€å¥åè§£æ, ãããã³ã°, åŸåã¹ã³ã¢ (æšå®ãšèšºæ), é確çéã¿ä»ã(IPTW), æšæºå (G-computation), åšèŸºæ§é ã¢ãã«(MSM), æéäŸåæ§äº€çµ¡ |
| C27 | 第27åïŒèªç¶å®éšã峿¹ã«ã€ãããæºå®éšçãã¶ã€ã³ã®åŒ·åããŒã« | å·®åã®å·®åæ³(DID), å¹³è¡ãã¬ã³ãä»®å®, ååž°äžé£ç¶ãã¶ã€ã³(RDD), ã©ã³ãã³ã°å€æ°, ã«ãããªã, é£ç¶æ§ã®ä»®å®, Sharp RDD vs Fuzzy RDD, æäœå€æ°æ³(IV), å çæ§, æäœå€æ°ã®3èŠä»¶ (é¢é£æ§, é€å€å¶çŽ, å調æ§), 屿çå¹³ååŠçœ®å¹æ(LATE) |
| C28 | 第28åïŒAIãšå ææšè«ã®ããªãŒã ããŒã ãçµæãã | Double Machine Learning(DML), Meta-Learners (S/T/X-learner), ç°è³ªæ§åŠçœ®å¹æ(CATE), ã¢ãããªããã¢ããªã³ã°, å æãã©ã¬ã¹ã (Causal Forest) |
| C29 | 第29åïŒããŒã¿ããå æã®å°å³ãæãåºããå æçºèŠã®ã¢ãããŒã | å æçºèŠ (Causal Discovery), å æçãã«ã³ãæ¡ä»¶ãšå¿ 宿§, ãæ§é åŠç¿ã: å¶çŽããŒã¹æ³ (PCã¢ã«ãŽãªãºã ), ã¹ã³ã¢ããŒã¹æ³, 颿°åå æã¢ãã« (LiNGAM), ãç¹å®å¹æã®æ€èšŒã: ã¡ã³ãã«ã©ã³ãã å(MR)ãšæ°Žå¹³çå€é¢çºçŸ |
C20-29
- 第20åïŒC20ïŒããŒã¿ãšéåŒã«èªãåãæè¡ãçµ±èšçæèãšå¯èŠåã®ç¬¬äžæ©
â èšè¿°çµ±èš (代衚å€, æ£åžåºŠ), æšæž¬çµ±èš (æ¯éå£ãšæšæ¬), ããŒã¿å¯èŠå (EDA, ãã¹ãã°ã©ã , ç®±ã²ãå³), 確çã®åºç€ (確ç倿°, æ¡ä»¶ä»ã確ç, ãã€ãºã®å®ç), 確çååž (æ£èŠååž, äºé ååž, ãã¢ãœã³ååž), 倧æ°ã®æ³å, äžå¿æ¥µéå®ç - 第21åïŒC21ïŒãå¶ç¶ããå¿
ç¶ãïŒããèŠæ¥µããç§åŠ â 仮説æ€å®ãšä¿¡é Œåºéã䜿ãããªã
â 仮説æ€å® (åž°ç¡ä»®èª¬, 察ç«ä»®èª¬, på€, æææ°Žæº), 第1çš®ã»ç¬¬2çš®ã®é誀, æ€åºå, ä¿¡é Œåºé, tæ€å® (察å¿ãã/ãªã), ã«ã€äºä¹æ€å®, 忣åæ(ANOVA), å€éæ¯èŒ - 第22åïŒC22ïŒããªãïŒãã«çããç§åŠãå ææšè«ãžã®æåŸ
ç¶
â çžé¢ãšå æ, 亀絡, 第3ã®å€æ°, æ§é çå æã¢ãã«(SCM), æåéå·¡åã°ã©ã(DAG), dåé¢, ããã¯ãã¢åºæº, åå®ä»®æ³, ããã³ã·ã£ã«ã¢ãŠãã«ã , å æå¹æ (ATE, ATT), èå¥å¯èœæ§ã®3å€§ä»®å® (亀æå¯èœæ§, äžè²«æ§, ããžãã£ããã£) - 第23åïŒC23ïŒæåŒ·ã®ãšããã³ã¹ãåµããèšåºç«åŠãšç ç©¶ãã¶ã€ã³ â ããŒã¿ãããçå®ããèŠæãèªæµ·è¡
â ç«åŠææš (眹æ£ç, æç ç, ãªã¹ã¯æ¯, ãªããºæ¯, å¯äžå±éº), ãã€ã¢ã¹ (éžæãã€ã¢ã¹, æ å ±ãã€ã¢ã¹), ç ç©¶ãã¶ã€ã³ (暪æç ç©¶, çäŸå¯Ÿç §ç ç©¶, ã³ããŒãç ç©¶, ã©ã³ãã 忝èŒè©Šéš/RCT), 蚺æç²ŸåºŠææš (æåºŠ, ç¹ç°åºŠ, ROCæ²ç·, AUC) - 第24åïŒC24ïŒæªæ¥ãäºæž¬ããçµ±èšã¢ããªã³ã°â ïŒååž°åæãã¹ã¿ãŒ
â ç·åœ¢éååž°åæ (æå°äºä¹æ³), ããžã¹ãã£ãã¯ååž°åæ (æå°€æ³), ãã¢ãœã³ååž°, ã¢ãã«èšºæ (æ®å·®åæ, å€éå ±ç·æ§/VIF), 倿°éžæ (ã¹ãããã¯ã€ãº, LASSO), ã¢ãã«è©äŸ¡ (決å®ä¿æ°, AIC, AUC) - 第25åïŒC25ïŒæªæ¥ãäºæž¬ããçµ±èšã¢ããªã³ã°â¡ïŒçåæéåæ
â æã¡åã (å³åŽ/å·ŠåŽ), çå颿°, ãã¶ãŒã颿°, Kaplan-Meieræ³, ãã°ã©ã³ã¯æ€å®, Coxæ¯äŸãã¶ãŒãã¢ãã«, æ¯äŸãã¶ãŒãæ§ã®ä»®å®, æéäŸåæ§å ±å€é - 第26åïŒC26ïŒãèŠããªãæµã亀絡ãè¯éºã«ãã°ããã¯ããã¯
â å±€å¥åè§£æ, ãããã³ã°, åŸåã¹ã³ã¢ (æšå®ãšèšºæ), é確çéã¿ä»ã(IPTW), æšæºå (G-computation), åšèŸºæ§é ã¢ãã«(MSM), æéäŸåæ§äº€çµ¡ - 第27åïŒC27ïŒèªç¶å®éšã峿¹ã«ã€ãããæºå®éšçãã¶ã€ã³ã®åŒ·åããŒã«
â å·®åã®å·®åæ³(DID), å¹³è¡ãã¬ã³ãä»®å®, ååž°äžé£ç¶ãã¶ã€ã³(RDD), ã©ã³ãã³ã°å€æ°, ã«ãããªã, é£ç¶æ§ã®ä»®å®, Sharp RDD vs Fuzzy RDD, æäœå€æ°æ³(IV), å çæ§, æäœå€æ°ã®3èŠä»¶ (é¢é£æ§, é€å€å¶çŽ, å調æ§), 屿çå¹³ååŠçœ®å¹æ(LATE) - 第28åïŒC28ïŒAIãšå ææšè«ã®ããªãŒã ããŒã ãçµæãã
â Double Machine Learning(DML), Meta-Learners (S/T/X-learner), ç°è³ªæ§åŠçœ®å¹æ(CATE), ã¢ãããªããã¢ããªã³ã°, å æãã©ã¬ã¹ã (Causal Forest) - 第29åïŒC29ïŒããŒã¿ããå æã®å°å³ãæãåºããå æçºèŠã®ã¢ãããŒã
â å æçºèŠ (Causal Discovery), å æçãã«ã³ãæ¡ä»¶ãšå¿ 宿§, ãæ§é åŠç¿ã: å¶çŽããŒã¹æ³ (PCã¢ã«ãŽãªãºã ), ã¹ã³ã¢ããŒã¹æ³, 颿°åå æã¢ãã« (LiNGAM), ãç¹å®å¹æã®æ€èšŒã: ã¡ã³ãã«ã©ã³ãã å(MR)ãšæ°Žå¹³çå€é¢çºçŸ
第IIIéšïŒã¢ã€ãã¢ãã«ã¿ãã«ãPythonããã°ã©ãã³ã°ã®åéº

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C30 | 第30åïŒå»çAI à Pythonéçºã®å°å³ãæã«å ¥ãããã2026幎ããŒãžã§ã³ã | Python, å»çAI, Jupyter Notebook, VS Code, ä»®æ³ç°å¢, Docker, Git, éçºãããŒå šäœå |
| C30.1 | Python, ããŒã¿ãµã€ãšã³ã¹, ã©ã€ãã©ãªãšã³ã·ã¹ãã , Jupyter, VS Code, ä»®æ³ç°å¢ã®åœ¹å² | |
| C30.2 | Jupyter Notebook, JupyterLab, ã»ã«æ§é , EDA, å¯èŠå, å®éšãã°, åçŸæ§ | |
| C30.3 | VS Code, ãããžã§ã¯ãæ§æ, srcãã©ã«ã, ãããã¬, ã¿ãŒããã«, Jupyter飿º, å®è£ ãã㌠| |
| C30.4 | ä»®æ³ç°å¢, venv, pip, requirements.txt, äŸåé¢ä¿ç®¡ç, ç°å¢æ±æé²æ¢ | |
| C30.5 | uv, conda, Miniconda, GPU, CUDA, Docker, ã³ã³ãã, æ¬çªããã〠| |
| C30.6 | 倿°, ããŒã¿å, ifæ, foræ, å å 衚èš, 颿°, ã¯ã©ã¹, äŸå€åŠç, ãã¡ã€ã«I/O | |
| C30.7 | Git, GitHub, ã³ããã, ãã©ã³ã, ãã«ãªã¯ãšã¹ã, ããŒãžã§ã³ç®¡ç, åçŸæ§, ããŒã éçº | |
| C31 | 第31åïŒããŒã¿åéºè ã®äžçš®ã®ç¥åšïŒNumPy, Pandas, Matplotlibã䜿ãããªã | NumPy (é åæäœ), Pandas (DataFrame, ããŒã¿èªã¿èŸŒã¿/å å·¥), Matplotlib & Seaborn (ããŒã¿å¯èŠå), Scikit-learn (ååŠç) |
| C31.1 | æ¬ æã¡ã«ããºã (MCAR/MAR/MNAR), å»åŠçãã€ã¢ã¹, åçŽä»£å ¥ vs å€éä»£å ¥(MICE), å€ã倿€ç¥ | |
| C31.2 | æ±ºå®æš, ã©ã³ãã ãã©ã¬ã¹ã, GBDT(XGBoost/LightGBM), SVM, ç·åœ¢ vs éç·åœ¢, èšåºã¹ã³ã¢äœæ | |
| C31.3 | æ£è å±€å¥å(ãã§ãã¿ã€ãã³ã°), K-Means, éå±€çã¯ã©ã¹ã¿ãªã³ã°, 次å åæž(PCA/t-SNE/UMAP), ç°åžžæ€ç¥ | |
| C31.4 | äžåè¡¡ããŒã¿æŠç¥(SMOTE), Precision-Recallæ²ç·, XAI(SHAP/LIME), ç¹åŸŽééèŠåºŠ, èšåºçä¿¡é Œæ§ | |
| C32 | 第32åïŒAIã®å¿èéšãèŠãïŒPyTorchã§ãã£ãŒãã©ãŒãã³ã°ãåããã | Tensor (ãã³ãœã«), èªå埮å (autograd), èšç®ã°ã©ã, `nn.Module`ã«ãããã¥ãŒã©ã«ãããã¯ãŒã¯ã®æ§ç¯, `forward`ã¡ãœãã |
| C33 | 第33åïŒAIãè³¢ãè²ãŠããåŠç¿ãµã€ã¯ã«ãïŒæå€±é¢æ°ããæé©åãŸã§ | æå€±é¢æ° (MSE, 亀差ãšã³ããããŒ), æé©åã¢ã«ãŽãªãºã (SGD, Adam), åŠç¿ç, ãããããåŠç¿, DataLoader, ãšãã㯠|
| C34 | 第34åïŒãäœã£ãŠåŠã¶ãAIã®ãç®ããåµãïŒç»åèªèã¢ãã«(CNN)å®è·µå ¥é | ç³ã¿èŸŒã¿å±€, ããŒãªã³ã°å±€, 掻æ§å颿°(ReLU), CNNã¢ãŒããã¯ã㣠(LeNet, VGG), ããŒã¿æ¡åŒµ, 転移åŠç¿ |
| C35 | 第35åïŒãäœã£ãŠåŠã¶ãAIã«ãæéããæããïŒæç³»åã¢ãã«(RNN/LSTM)å ¥é | RNN (ååž°åãã¥ãŒã©ã«ãããã¯ãŒã¯), åŸé æ¶å€±åé¡, LSTM, GRU, ã²ãŒãæ©æ§, ç³»åããŒã¿åŠç, èªç¶èšèªåŠçãžã®å¿çš |
| C36 | 第36åïŒãäœã£ãŠåŠã¶ãçŸä»£AIã®çè ïŒTransformerã¢ãã«åŸ¹åºè§£å | Self-Attention, Multi-Head Attention, Positional Encoding, Encoder-Decoderã¢ãã«, BERT, Vision Transformer(ViT) |
| C37 | 第37åïŒããªãã®AIãæåŒ·ã«ããïŒã¢ãã«è©äŸ¡ãšæ¹åã®å¿ é ãã¯ãã㯠| è©äŸ¡ææš (æ£è§£ç, é©åç, åçŸç, F1ã¹ã³ã¢, AUC), æ··åè¡å, éåŠç¿å¯Ÿç (æ£åå, Dropout), ãã€ããŒãã©ã¡ãŒã¿ãã¥ãŒãã³ã° |
| C38 | 第38åïŒè²ãŠãAIãå®äžçãžïŒã¢ãã«ã®ä¿åãšæšè«ïŒãããã€ïŒã®ç¬¬äžæ© | åŠç¿æžã¿ã¢ãã«ã®ã·ãªã¢ã©ã€ãº (`.pt`, `.pth`), æšè«ã¢ãŒã (`model.eval()`), ONNX, FastAPIã«ããAPIåã®åºç€ |
| C39 | 第39åïŒäžæµã®ã³ãŒããæžãïŒåçŸæ§ãšä¿å®æ§ãé«ããPythonããã®äœæ³ | ãªããžã§ã¯ãæåããã°ã©ãã³ã° (ã¯ã©ã¹), åãã³ã, docstring, ãã¹ãã³ãŒã(pytest), ã³ãŒããã©ãŒããã¿ (Black, isort), Lint (flake8) |
| C39.1 | SQL, SELECT, JOIN, Window颿°, BigQuery, MIMIC-IV | |
| C39.2 | Rèšèª, rpy2, ggplot2, çåæéåæ, å³å¯ãªçµ±èš | |
| C39.3 | JavaScript, TypeScript, React, D3.js, Chart.js, ããã³ããšã³ã, FastAPI飿º | |
| C39.4 | C++, çµã¿èŸŒã¿ã·ã¹ãã , é«éç»ååŠç, ãšããžAI, å»çæ©åšãœãããŠã§ã¢ | |
| C39.5 | Bash, Shell, Linux, ã¯ã©ãŠããµãŒããŒ(AWS/GCP), Cron, èªåå | |
| C39.6 | FHIR, HL7, JSON, API飿º, å»çæ å ±äº€æ, ã€ã³ã¿ãŒãªãã©ããªã㣠| |
| C39.7 | å¿åå, å å·¥, k-å¿åå, å·®åãã©ã€ãã·ãŒ, ã»ãã¥ã¢ã³ã³ãã¥ããŒã·ã§ã³, å人æ å ±ä¿è· |
C30-39
- 第30åïŒC30ïŒå»çAI à Pythonéçºã®å°å³ãæã«å
¥ãããã2026幎ããŒãžã§ã³ã
â Python, å»çAI, Jupyter Notebook, VS Code, ä»®æ³ç°å¢, Docker, Git, éçºãããŒå šäœå- 第30.1åïŒC30.1ïŒãªãPythonãªã®ãïŒå»çAIãšéçºããŒã«ã®å
šäœå
â Python, ããŒã¿ãµã€ãšã³ã¹, ã©ã€ãã©ãªãšã³ã·ã¹ãã , Jupyter, VS Code, ä»®æ³ç°å¢ã®åœ¹å² - 第30.2åïŒC30.2ïŒJupyter Notebookã§å§ãããããžã¿ã«å®éšããŒããå
¥é
â Jupyter Notebook, JupyterLab, ã»ã«æ§é , EDA, å¯èŠå, å®éšãã°, åçŸæ§ - 第30.3åïŒC30.3ïŒVS Codeã§ãæ¬çªãæèãããPythonéçºã«ä¹ãæããã
â VS Code, ãããžã§ã¯ãæ§æ, srcãã©ã«ã, ãããã¬, ã¿ãŒããã«, Jupyter飿º, å®è£ ãã㌠- 第30.4åïŒC30.4ïŒä»®æ³ç°å¢å
¥éââvenvã§ãæž
æœéããå®ã
â ä»®æ³ç°å¢, venv, pip, requirements.txt, äŸåé¢ä¿ç®¡ç, ç°å¢æ±æé²æ¢ - 第30.5åïŒC30.5ïŒuvã»condaã»Dockerââçºå±ããŒã«ã§æ¬çªç°å¢ãŸã§èŠæ®ãã
â uv, conda, Miniconda, GPU, CUDA, Docker, ã³ã³ãã, æ¬çªããã〠- 第30.6åïŒC30.6ïŒå»çè
ã®ããã®Pythonææ³ã·ã§ãŒãã³ãŒã¹
â 倿°, ããŒã¿å, ifæ, foræ, å å 衚èš, 颿°, ã¯ã©ã¹, äŸå€åŠç, ãã¡ã€ã«I/O - 第30.7åïŒC30.7ïŒGitãšGitHubã§ãæãæ»ãããè§£æç°å¢ãäœã
â Git, GitHub, ã³ããã, ãã©ã³ã, ãã«ãªã¯ãšã¹ã, ããŒãžã§ã³ç®¡ç, åçŸæ§, ããŒã éçº
- 第30.1åïŒC30.1ïŒãªãPythonãªã®ãïŒå»çAIãšéçºããŒã«ã®å
šäœå
- 第31åïŒC31ïŒããŒã¿åéºè
ã®äžçš®ã®ç¥åšïŒNumPy, Pandas, Matplotlibã䜿ãããªã
â NumPy (é åæäœ), Pandas (DataFrame, ããŒã¿èªã¿èŸŒã¿/å å·¥), Matplotlib & Seaborn (ããŒã¿å¯èŠå), Scikit-learn (ååŠç)- 第31.1åïŒC31.1ïŒãååŠçæŠç¥ãRWDã®ãæ±ãããå»åŠçã«æ£ããåŠçãã
â æ¬ æã¡ã«ããºã (MCAR/MAR/MNAR), å»åŠçãã€ã¢ã¹, åçŽä»£å ¥ vs å€éä»£å ¥(MICE), å€ã倿€ç¥ - 第31.2åïŒC31.2ïŒãäºæž¬ã¢ãã«éžæãããã¯ã€ãããã¯ã¹ããã粟床ããïŒã¢ã«ãŽãªãºã éžå®ã®çŸ
éç€
â æ±ºå®æš, ã©ã³ãã ãã©ã¬ã¹ã, GBDT(XGBoost/LightGBM), SVM, ç·åœ¢ vs éç·åœ¢, èšåºã¹ã³ã¢äœæ - 第31.3åïŒC31.3ïŒããã¿ãŒã³çºèŠãæ£è§£ã®ãªãããŒã¿ãããæ£è
åããæµ®ãã³äžããããïŒæåž«ãªãåŠç¿ïŒ
â æ£è å±€å¥å(ãã§ãã¿ã€ãã³ã°), K-Means, éå±€çã¯ã©ã¹ã¿ãªã³ã°, 次å åæž(PCA/t-SNE/UMAP), ç°åžžæ€ç¥ - 第31.4åïŒC31.4ïŒãä¿¡é Œæ§ã®èšèšãäžåè¡¡ããŒã¿ãšèª¬æå¯èœæ§(XAI)ã§ã䜿ããAIãã«ãã
â äžåè¡¡ããŒã¿æŠç¥(SMOTE), Precision-Recallæ²ç·, XAI(SHAP/LIME), ç¹åŸŽééèŠåºŠ, èšåºçä¿¡é Œæ§
- 第31.1åïŒC31.1ïŒãååŠçæŠç¥ãRWDã®ãæ±ãããå»åŠçã«æ£ããåŠçãã
- 第32åïŒC32ïŒAIã®å¿èéšãèŠãïŒPyTorchã§ãã£ãŒãã©ãŒãã³ã°ãåããã
â Tensor (ãã³ãœã«), èªå埮å (autograd), èšç®ã°ã©ã, `nn.Module`ã«ãããã¥ãŒã©ã«ãããã¯ãŒã¯ã®æ§ç¯, `forward`ã¡ãœãã - 第33åïŒC33ïŒAIãè³¢ãè²ãŠããåŠç¿ãµã€ã¯ã«ãïŒæå€±é¢æ°ããæé©åãŸã§
â æå€±é¢æ° (MSE, 亀差ãšã³ããããŒ), æé©åã¢ã«ãŽãªãºã (SGD, Adam), åŠç¿ç, ãããããåŠç¿, DataLoader, ãšãã㯠- 第34åïŒC34ïŒãäœã£ãŠåŠã¶ãAIã®ãç®ããåµãïŒç»åèªèã¢ãã«(CNN)å®è·µå
¥é
â ç³ã¿èŸŒã¿å±€, ããŒãªã³ã°å±€, 掻æ§å颿°(ReLU), CNNã¢ãŒããã¯ã㣠(LeNet, VGG), ããŒã¿æ¡åŒµ, 転移åŠç¿ - 第35åïŒC35ïŒãäœã£ãŠåŠã¶ãAIã«ãæéããæããïŒæç³»åã¢ãã«(RNN/LSTM)å
¥é
â RNN (ååž°åãã¥ãŒã©ã«ãããã¯ãŒã¯), åŸé æ¶å€±åé¡, LSTM, GRU, ã²ãŒãæ©æ§, ç³»åããŒã¿åŠç, èªç¶èšèªåŠçãžã®å¿çš - 第36åïŒC36ïŒãäœã£ãŠåŠã¶ãçŸä»£AIã®çè
ïŒTransformerã¢ãã«åŸ¹åºè§£å
â Self-Attention, Multi-Head Attention, Positional Encoding, Encoder-Decoderã¢ãã«, BERT, Vision Transformer(ViT) - 第37åïŒC37ïŒããªãã®AIãæåŒ·ã«ããïŒã¢ãã«è©äŸ¡ãšæ¹åã®å¿
é ãã¯ããã¯
â è©äŸ¡ææš (æ£è§£ç, é©åç, åçŸç, F1ã¹ã³ã¢, AUC), æ··åè¡å, éåŠç¿å¯Ÿç (æ£åå, Dropout), ãã€ããŒãã©ã¡ãŒã¿ãã¥ãŒãã³ã° - 第38åïŒC38ïŒè²ãŠãAIãå®äžçãžïŒã¢ãã«ã®ä¿åãšæšè«ïŒãããã€ïŒã®ç¬¬äžæ©
â åŠç¿æžã¿ã¢ãã«ã®ã·ãªã¢ã©ã€ãº (`.pt`, `.pth`), æšè«ã¢ãŒã (`model.eval()`), ONNX, FastAPIã«ããAPIåã®åºç€ - 第39åïŒC39ïŒäžæµã®ã³ãŒããæžãïŒåçŸæ§ãšä¿å®æ§ãé«ããPythonããã®äœæ³
â ãªããžã§ã¯ãæåããã°ã©ãã³ã° (ã¯ã©ã¹), åãã³ã, docstring, ãã¹ãã³ãŒã(pytest), ã³ãŒããã©ãŒããã¿ (Black, isort), Lint (flake8)- 第39.1åïŒC39.1ïŒããŒã¿ã®æºæ³ãæãèµ·ããïŒé»åã«ã«ãæ»ç¥ã®ããã®SQLå
¥é
â SQL, SELECT, JOIN, Window颿°, BigQuery, MIMIC-IV - 第39.2åïŒC39.2ïŒçµ±èšè§£æã®ãæš©åšãã峿¹ã«ïŒPythonããæãRèšèªé£æº
â Rèšèª, rpy2, ggplot2, çåæéåæ, å³å¯ãªçµ±èš - 第39.3åïŒC39.3ïŒAIãã䜿ããã圢ã«ïŒJavaScriptãšReactã§åµãå»çUI
â JavaScript, TypeScript, React, D3.js, Chart.js, ããã³ããšã³ã, FastAPI飿º - 第39.4åïŒC39.4ïŒå»çæ©åšã®æ·±æ·µãžïŒC++ã§æãé«éåŠçãšçµã¿èŸŒã¿éçº
â C++, çµã¿èŸŒã¿ã·ã¹ãã , é«éç»ååŠç, ãšããžAI, å»çæ©åšãœãããŠã§ã¢ - 第39.5åïŒC39.5ïŒãµãŒããŒç®¡çã®åžä»€å¡ïŒBashã·ã§ã«ã§èªååãæ¥µãã
â Bash, Shell, Linux, ã¯ã©ãŠããµãŒããŒ(AWS/GCP), Cron, èªåå - 第39.6åïŒC39.6ïŒç
é¢ã·ã¹ãã ãšäŒè©±ãããïŒFHIR & HL7ããã°ã©ãã³ã°
â FHIR, HL7, JSON, API飿º, å»çæ å ±äº€æ, ã€ã³ã¿ãŒãªãã©ããªã㣠- 第39.7åïŒC39.7ïŒæ£è
ããŒã¿ãéå£ã«å®ãïŒãã©ã€ãã·ãŒãšã³ãžãã¢ãªã³ã°ã®å®è£
â å¿åå, å å·¥, k-å¿åå, å·®åãã©ã€ãã·ãŒ, ã»ãã¥ã¢ã³ã³ãã¥ããŒã·ã§ã³, å人æ å ±ä¿è·
- 第39.1åïŒC39.1ïŒããŒã¿ã®æºæ³ãæãèµ·ããïŒé»åã«ã«ãæ»ç¥ã®ããã®SQLå
¥é
第IVéšïŒAI㮔箔ãè²ãŠããå»çç»åè§£æã®æåç·

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C40 | 第40åïŒAIã®ãç®ãã§èŠãäžçãå»çç»åã®åºç€ç¥è | å»çç»åã®ã¢ããªã㣠(CT, MRI, Xç·, ç ç), AIã®ã¿ã¹ã¯ (åé¡, ç©äœæ€åº, ã»ã°ã¡ã³ããŒã·ã§ã³) |
| C41 | 第41åïŒAIã®ããã®ãçŸé£åŠããæé«ã®åŠç¿ããããç»åååŠç | ç»åæ£èŠå (èŒåºŠå€), ãŠã£ã³ããŠã€ã³ã° (CTå€), ããŒã¿æ¡åŒµ (Augmentation), ã¢ãããŒã·ã§ã³ (ããŠã³ãã£ã³ã°ããã¯ã¹, ã»ã°ã¡ã³ããŒã·ã§ã³ãã¹ã¯), ã¢ãããŒã·ã§ã³ããŒã« (Labelbox, ITK-SNAP), å質管ç |
| C42 | 第42åïŒç»åèªèã®çéãCNNã¢ãŒããã¯ãã£åŸ¹åºè§£å | ç³ã¿èŸŒã¿å±€, ããŒãªã³ã°å±€, 掻æ§å颿°(ReLU), CNNã¢ãŒããã¯ãã£ã®é²å (AlexNet, VGG, ResNet, DenseNet), ã¹ãããæ¥ç¶ |
| C43 | 第43åïŒå·šäººã®è©ã«ä¹ãåŠç¿æ³ã転移åŠç¿ã䜿ãããªã | ImageNetäºååŠç¿ã¢ãã«, ãã¡ã€ã³ãã¥ãŒãã³ã° (å šå±€ vs. äžéš), ãã¡ã€ã³é©å¿, å°ãªãããŒã¿ã§ã®åŠç¿æŠç¥ |
| C44 | 第44åïŒç å€ã®èŒªéããã¯ã»ã«åäœã§æãããã»ã°ã¡ã³ããŒã·ã§ã³æè¡ | ã»ãã³ãã£ãã¯ã»ã°ã¡ã³ããŒã·ã§ã³, U-Net (ãšã³ã³ãŒã/ãã³ãŒãæ§é ), ã¹ãããæ¥ç¶, 3D U-Net, è©äŸ¡ææš (Diceä¿æ°, IoU) |
| C45 | 第45åïŒç»åã®äžãããå®ç©ããæ¢ãåºããç©äœæ€åºæè¡ | ç©äœæ€åº, ããŠã³ãã£ã³ã°ããã¯ã¹, one-stage (YOLO) vs. two-stage (Faster R-CNN) detector, ã¢ã³ã«ãŒããã¯ã¹, è©äŸ¡ææš(mAP) |
| C46 | 第46åïŒ2Dãã3DãžïŒCT/MRIãç«äœçã«è§£æããæè¡ | 3D-CNN, 3DããŒã¿ãã³ããªã³ã° (NIfTI, NumPy), èšç®ãªãœãŒã¹ã®èª²é¡ (VRAM), ãããããŒã¹åŠç¿, 2.5Dã¢ãããŒã |
| C47 | 第47åïŒAIã®ãæèããå¯èŠåããXAIïŒèª¬æå¯èœãªAIïŒãå€æã®æ ¹æ ãæ¢ã | 説æå¯èœãªAI(XAI), Grad-CAM, LIME, SHAP, å€ææ ¹æ ã®å¯èŠå, èšåºçåŠ¥åœæ§ã®è©äŸ¡, ä¿¡é Œæ§ |
| C48 | 第48åïŒç»åèªèã®æ°æä»£ãVision Transformer (ViT) | Vision Transformer(ViT), ç»åã®ãããåãšããŒã¯ã³å, Self-Attention, Positional Embedding, CNNãšã®æ¯èŒ, 倧åçæèã®ç²åŸ |
| C49 | 第49åïŒç»åãšèšèã§èšºæããããã«ãã¢ãŒãã«AI | ãã«ãã¢ãŒãã«AI, ç»åãšèšåºããã¹ãã®çµ±å, CLIP (å¯Ÿç §åŠç¿), å ±ååã蟌ã¿ç©ºé, ãŒãã·ã§ããåé¡, 蚺æã¬ããŒãèªåçæ |
C40-49
- 第40åïŒC40ïŒAIã®ãç®ãã§èŠãäžçãå»çç»åã®åºç€ç¥è
â å»çç»åã®ã¢ããªã㣠(CT, MRI, Xç·, ç ç), AIã®ã¿ã¹ã¯ (åé¡, ç©äœæ€åº, ã»ã°ã¡ã³ããŒã·ã§ã³) - 第41åïŒC41ïŒAIã®ããã®ãçŸé£åŠããæé«ã®åŠç¿ããããç»åååŠç
â ç»åæ£èŠå (èŒåºŠå€), ãŠã£ã³ããŠã€ã³ã° (CTå€), ããŒã¿æ¡åŒµ (Augmentation), ã¢ãããŒã·ã§ã³ (ããŠã³ãã£ã³ã°ããã¯ã¹, ã»ã°ã¡ã³ããŒã·ã§ã³ãã¹ã¯), ã¢ãããŒã·ã§ã³ããŒã« (Labelbox, ITK-SNAP), å質管ç - 第42åïŒC42ïŒç»åèªèã®çéãCNNã¢ãŒããã¯ãã£åŸ¹åºè§£å
â ç³ã¿èŸŒã¿å±€, ããŒãªã³ã°å±€, 掻æ§å颿°(ReLU), CNNã¢ãŒããã¯ãã£ã®é²å (AlexNet, VGG, ResNet, DenseNet), ã¹ãããæ¥ç¶ - 第43åïŒC43ïŒå·šäººã®è©ã«ä¹ãåŠç¿æ³ã転移åŠç¿ã䜿ãããªã
â ImageNetäºååŠç¿ã¢ãã«, ãã¡ã€ã³ãã¥ãŒãã³ã° (å šå±€ vs. äžéš), ãã¡ã€ã³é©å¿, å°ãªãããŒã¿ã§ã®åŠç¿æŠç¥ - 第44åïŒC44ïŒç
å€ã®èŒªéããã¯ã»ã«åäœã§æãããã»ã°ã¡ã³ããŒã·ã§ã³æè¡
â ã»ãã³ãã£ãã¯ã»ã°ã¡ã³ããŒã·ã§ã³, U-Net (ãšã³ã³ãŒã/ãã³ãŒãæ§é ), ã¹ãããæ¥ç¶, 3D U-Net, è©äŸ¡ææš (Diceä¿æ°, IoU) - 第45åïŒC45ïŒç»åã®äžãããå®ç©ããæ¢ãåºããç©äœæ€åºæè¡
â ç©äœæ€åº, ããŠã³ãã£ã³ã°ããã¯ã¹, one-stage (YOLO) vs. two-stage (Faster R-CNN) detector, ã¢ã³ã«ãŒããã¯ã¹, è©äŸ¡ææš(mAP) - 第46åïŒC46ïŒ2Dãã3DãžïŒCT/MRIãç«äœçã«è§£æããæè¡
â 3D-CNN, 3DããŒã¿ãã³ããªã³ã° (NIfTI, NumPy), èšç®ãªãœãŒã¹ã®èª²é¡ (VRAM), ãããããŒã¹åŠç¿, 2.5Dã¢ãããŒã - 第47åïŒC47ïŒAIã®ãæèããå¯èŠåããXAIïŒèª¬æå¯èœãªAIïŒãå€æã®æ ¹æ ãæ¢ã
â 説æå¯èœãªAI(XAI), Grad-CAM, LIME, SHAP, å€ææ ¹æ ã®å¯èŠå, èšåºçåŠ¥åœæ§ã®è©äŸ¡, ä¿¡é Œæ§ - 第48åïŒC48ïŒç»åèªèã®æ°æä»£ãVision Transformer (ViT)
â Vision Transformer(ViT), ç»åã®ãããåãšããŒã¯ã³å, Self-Attention, Positional Embedding, CNNãšã®æ¯èŒ, 倧åçæèã®ç²åŸ - 第49åïŒC49ïŒç»åãšèšèã§èšºæããããã«ãã¢ãŒãã«AI
â ãã«ãã¢ãŒãã«AI, ç»åãšèšåºããã¹ãã®çµ±å, CLIP (å¯Ÿç §åŠç¿), å ±ååã蟌ã¿ç©ºé, ãŒãã·ã§ããåé¡, 蚺æã¬ããŒãèªåçæ
第VéšïŒAIã«”èšèã®å¿”ãæãããèšåºããã¹ããã€ãã³ã°

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C50 | 第50åïŒã«ã«ãã¯å®ã®å±±ïŒå»çèªç¶èšèªåŠçïŒNLPïŒã®å¯èœæ§ãšåºç€æè¡ | éæ§é åããŒã¿, ããŒã¯ã³å, 圢æ çŽ è§£æ, n-gram, Bag-of-Words, TF-IDF, åæ£è¡šçŸ, Word2Vec, FastText |
| C51 | 第51åïŒAIãæèãèªãç§å¯ãBERTãšTransformerã®é©æ° | Transformer, Self-Attention, Multi-Head Attention, Positional Encoding, äºååŠç¿ãšãã¡ã€ã³ãã¥ãŒãã³ã°, BERT, ãšã³ã³ãŒã |
| C52 | 第52åïŒããã¹ãããæ å ±ãèªåæœåºâ ïŒåºæè¡šçŸæœåºïŒNERïŒã§å®ãæ¢ã | åºæè¡šçŸæœåº(NER), IOB/BIOæ³, çŸæ£å, è¬å€å, æ€æ»å€ã®æœåº, ã¢ãããŒã·ã§ã³, ã¹ãã³ããŒã¹NER, å»çç¹åNER |
| C53 | 第53åïŒããã¹ãããæ å ±ãèªåæœåºâ¡ïŒé¢ä¿æœåºã§æ å ±ã®ç¹ãããè§£ãæãã | é¢ä¿æœåº(RE), ãšã³ãã£ãã£ãã¢, è¬å€ãšå¯äœçš, çŸæ£ãšçç¶, äŸåæ§æè§£æ, ãã¿ãŒã³ããŒã¹, æ©æ¢°åŠç¿ããŒã¹ |
| C54 | 第54åïŒAIã«ããã«ã«ãã®èªååé¡ãšã©ããªã³ã°ãICDã³ãŒãã£ã³ã°ãæ¯æŽãã | ããã¹ãåé¡, ãã«ãã¯ã©ã¹åé¡, ãã«ãã©ãã«åé¡, ICDã³ãŒãã£ã³ã°æ¯æŽ, æå®³äºè±¡å ±åã®åé¡, éé¢ãµããª, éå±€çåé¡ |
| C55 | 第55åïŒé·æãç¬æã«èŠçŽïŒè«æã»ã«ã«ãã®èªåèŠçŽè¡ | èªåèŠçŽ, æœåºåèŠçŽ, çæåèŠçŽ, BART, T5, Pointer-Generator Networks, è©äŸ¡ææš (ROUGE, BLEU) |
| C56 | 第56åïŒAIãšå¯Ÿè©±ããæªæ¥ãå»çQAãšãã£ãããããã®æ§ç¯ | 察話ã·ã¹ãã , ãã£ããããã, 質åå¿ç(QA), æ€çŽ¢ããŒã¹QA, çæããŒã¹QA, RAG (Retrieval-Augmented Generation), æ£è æè² |
| C57 | 第57åïŒGPT-4ãåãæãå»çèšèªã¢ãã«ã®æ°æ¬¡å | å€§èŠæš¡èšèªã¢ãã«(LLM), åºç€ã¢ãã«, GPT, Med-PaLM, ããã³ãããšã³ãžãã¢ãªã³ã°, Few-shot learning, Chain-of-Thought |
| C58 | 第58åïŒAIã®ãç¥ã£ããã¶ãããèŠæãïŒãã«ã·ããŒã·ã§ã³ãšå«ççèª²é¡ | ãã«ã·ããŒã·ã§ã³, äºå®ã«åºã¥ããªãçæ, ãã€ã¢ã¹ (瀟äŒç, ããŒã¿), å ¬å¹³æ§, ãã©ã€ãã·ãŒä¿è·, å人æ å ±, å«ç (ELSI) |
| C59 | 第59åïŒãå®è·µãHugging Faceã§å»çNLPãã€ãã©ã€ã³ãæ§ç¯ãã | Hugging Face, Transformersã©ã€ãã©ãª, ãã€ãã©ã€ã³, å»çããŒã¿ã»ãã (MIMIC, PubMed), ãã¡ã€ã³ãã¥ãŒãã³ã°å®è·µ, ããŒã¯ãã€ã¶, ã¢ãã«ãã |
C50-59
- 第50åïŒC50ïŒã«ã«ãã¯å®ã®å±±ïŒå»çèªç¶èšèªåŠçïŒNLPïŒã®å¯èœæ§ãšåºç€æè¡
â éæ§é åããŒã¿, ããŒã¯ã³å, 圢æ çŽ è§£æ, n-gram, Bag-of-Words, TF-IDF, åæ£è¡šçŸ, Word2Vec, FastText - 第51åïŒC51ïŒAIãæèãèªãç§å¯ãBERTãšTransformerã®é©æ°
â Transformer, Self-Attention, Multi-Head Attention, Positional Encoding, äºååŠç¿ãšãã¡ã€ã³ãã¥ãŒãã³ã°, BERT, ãšã³ã³ãŒã - 第52åïŒC52ïŒããã¹ãããæ
å ±ãèªåæœåºâ ïŒåºæè¡šçŸæœåºïŒNERïŒã§å®ãæ¢ã
â åºæè¡šçŸæœåº(NER), IOB/BIOæ³, çŸæ£å, è¬å€å, æ€æ»å€ã®æœåº, ã¢ãããŒã·ã§ã³, ã¹ãã³ããŒã¹NER, å»çç¹åNER - 第53åïŒC53ïŒããã¹ãããæ
å ±ãèªåæœåºâ¡ïŒé¢ä¿æœåºã§æ
å ±ã®ç¹ãããè§£ãæãã
â é¢ä¿æœåº(RE), ãšã³ãã£ãã£ãã¢, è¬å€ãšå¯äœçš, çŸæ£ãšçç¶, äŸåæ§æè§£æ, ãã¿ãŒã³ããŒã¹, æ©æ¢°åŠç¿ããŒã¹ - 第54åïŒC54ïŒAIã«ããã«ã«ãã®èªååé¡ãšã©ããªã³ã°ãICDã³ãŒãã£ã³ã°ãæ¯æŽãã
â ããã¹ãåé¡, ãã«ãã¯ã©ã¹åé¡, ãã«ãã©ãã«åé¡, ICDã³ãŒãã£ã³ã°æ¯æŽ, æå®³äºè±¡å ±åã®åé¡, éé¢ãµããª, éå±€çåé¡ - 第55åïŒC55ïŒé·æãç¬æã«èŠçŽïŒè«æã»ã«ã«ãã®èªåèŠçŽè¡
â èªåèŠçŽ, æœåºåèŠçŽ, çæåèŠçŽ, BART, T5, Pointer-Generator Networks, è©äŸ¡ææš (ROUGE, BLEU) - 第56åïŒC56ïŒAIãšå¯Ÿè©±ããæªæ¥ãå»çQAãšãã£ãããããã®æ§ç¯
â 察話ã·ã¹ãã , ãã£ããããã, 質åå¿ç(QA), æ€çŽ¢ããŒã¹QA, çæããŒã¹QA, RAG (Retrieval-Augmented Generation), æ£è æè² - 第57åïŒC57ïŒGPT-4ãåãæãå»çèšèªã¢ãã«ã®æ°æ¬¡å
â å€§èŠæš¡èšèªã¢ãã«(LLM), åºç€ã¢ãã«, GPT, Med-PaLM, ããã³ãããšã³ãžãã¢ãªã³ã°, Few-shot learning, Chain-of-Thought - 第58åïŒC58ïŒAIã®ãç¥ã£ããã¶ãããèŠæãïŒãã«ã·ããŒã·ã§ã³ãšå«çç課é¡
â ãã«ã·ããŒã·ã§ã³, äºå®ã«åºã¥ããªãçæ, ãã€ã¢ã¹ (瀟äŒç, ããŒã¿), å ¬å¹³æ§, ãã©ã€ãã·ãŒä¿è·, å人æ å ±, å«ç (ELSI) - 第59åïŒC59ïŒãå®è·µãHugging Faceã§å»çNLPãã€ãã©ã€ã³ãæ§ç¯ãã
â Hugging Face, Transformersã©ã€ãã©ãª, ãã€ãã©ã€ã³, å»çããŒã¿ã»ãã (MIMIC, PubMed), ãã¡ã€ã³ãã¥ãŒãã³ã°å®è·µ, ããŒã¯ãã€ã¶, ã¢ãã«ãã
第VIéšïŒæªæ¥ãèªãã¿ã€ã ãã·ã³ãçäœä¿¡å·è§£æ

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C60 | 第60åïŒãã€ã¿ã«ãµã€ã³ã«é ãããç©èªãå»çæç³»åããŒã¿ã®äŸ¡å€ãšå¯èœæ§ | çäœä¿¡å· (ECG, EEG), ICUããŒã¿, é»åã«ã«ã(EHR), ãŠã§ã¢ã©ãã«ããã€ã¹, ã¿ã¹ã¯ (äºæž¬, ç°åžžæ€ç¥, åé¡), æç³»åããŒã¿ã®ç¹åŸŽ |
| C61 | 第61åïŒãã€ãºã®æµ·ããå®ãæããå»çæç³»åããŒã¿ã®ååŠçæè¡ | äžèŠåãµã³ããªã³ã° (è£é, ãªãµã³ããªã³ã°), æ¬ æå€è£å®, ãã€ãºé€å» (ç§»åå¹³å, ãã£ã«ã¿ãªã³ã°), ãŠã£ã³ããŠã€ã³ã°, ç¹åŸŽéãšã³ãžãã¢ãªã³ã° |
| C62 | 第62åïŒæ·±å±€åŠç¿ã ããããªãïŒçµ±èšçæç³»åã¢ãã«ã®å¡æºãšå®è·µ | å®åžžæ§, èªå·±çžé¢(ACF/PACF), ARIMAã¢ãã«, ç¶æ 空éã¢ãã«, ã«ã«ãã³ãã£ã«ã¿, ããŒã¹ã©ã€ã³ã¢ãã«ãšããŠã®éèŠæ§ |
| C63 | 第63åïŒéå»ãèšæ¶ããAIãRNNãšLSTM/GRUã®ã¡ã«ããºã | ååž°çæ§é , é·æäŸåæ§, åŸé æ¶å€±ã»ççºåé¡, LSTM, GRU, ã²ãŒãæ©æ§ (å ¥å, å¿åŽ, åºå) |
| C64 | 第64åïŒæªæ¥ãäºæž¬ãã翻蚳æ©ãSeq2Seqãšã¢ãã³ã·ã§ã³æ©æ§ | ãšã³ã³ãŒãã»ãã³ãŒãã¢ãã«, Seq2Seq, æèãã¯ãã«, ã¢ãã³ã·ã§ã³æ©æ§, æç³»åäºæž¬, æ©æ¢°ç¿»èš³ã®ã¢ãããžãŒ |
| C65 | 第65åïŒç»åã ããããªãïŒæç³»åè§£æã«ãããCNNã®é ããå | 1D-CNN, 屿çãã¿ãŒã³ã®æœåº, å æçç³ã¿èŸŒã¿, æ¡åŒµç³ã¿èŸŒã¿, å容é, WaveNet, TCN (Temporal Convolutional Networks) |
| C66 | 第66åïŒæç³»åè§£æã®é©åœå ãTransformerã®åšåãšå¿çš | Self-Attentionã«ããé·æäŸåæ§ã®ææ, Positional Encoding, èšç®å¹çã®èª²é¡, ææ°ã¢ãã« (Informer, PatchTST), å»çå¿çš |
| C66.1 | State Space Models (SSM), Mamba architecture, Samba (Hybrid), èšç®å¹çãšç²ŸåºŠã®ãã¬ãŒããªã, mamba-ssm ã©ã€ãã©ãªå®è£ | |
| C67 | 第67åïŒè€æ°ã®æ å ±ãçµ±åããããã«ãã¢ãŒãã«æç³»åè§£æ | ãã«ãã¢ãŒãã«åŠç¿, éçããŒã¿ (æ£è èæ¯) ãšåçããŒã¿ (æç³»å) ã®çµ±å, ããŒã¿ãã¥ãŒãžã§ã³ (æ©æ/åŸæ), èšåºããã¹ããšã®é£æº |
| C68 | 第68åïŒAIã®æèããã»ã¹ãèŠããæç³»åã¢ãã«ã®XAIïŒèª¬æå¯èœãªAIïŒ | 説æå¯èœãªAI (XAI), ã¢ãã³ã·ã§ã³ãããã®å¯èŠå, ç¹åŸŽééèŠåºŠ, LIME, SHAP, å€ææ ¹æ ã®ç¹å®, èšåºçä¿¡é Œæ§ |
| C69 | 第69åïŒãå®è·µãICUããŒã¿ã§æè¡çã®å åãæããäºæž¬ã¢ãã«ãæ§ç¯ãã | ICUããŒã¿ã»ãã (MIMIC-IV), ããŒã¿ååŠçãã€ãã©ã€ã³, ã¢ãã«æ§ç¯ãšèšç·Ž, æ§èœè©äŸ¡ (AUC, F1ã¹ã³ã¢), èšåºã·ããªãªã§ã®è©äŸ¡, çµæã®è§£é |
C60-69
- 第60åïŒC60ïŒãã€ã¿ã«ãµã€ã³ã«é ãããç©èªãå»çæç³»åããŒã¿ã®äŸ¡å€ãšå¯èœæ§
â çäœä¿¡å· (ECG, EEG), ICUããŒã¿, é»åã«ã«ã(EHR), ãŠã§ã¢ã©ãã«ããã€ã¹, ã¿ã¹ã¯ (äºæž¬, ç°åžžæ€ç¥, åé¡), æç³»åããŒã¿ã®ç¹åŸŽ - 第61åïŒC61ïŒãã€ãºã®æµ·ããå®ãæããå»çæç³»åããŒã¿ã®ååŠçæè¡
â äžèŠåãµã³ããªã³ã° (è£é, ãªãµã³ããªã³ã°), æ¬ æå€è£å®, ãã€ãºé€å» (ç§»åå¹³å, ãã£ã«ã¿ãªã³ã°), ãŠã£ã³ããŠã€ã³ã°, ç¹åŸŽéãšã³ãžãã¢ãªã³ã° - 第62åïŒC62ïŒæ·±å±€åŠç¿ã ããããªãïŒçµ±èšçæç³»åã¢ãã«ã®å¡æºãšå®è·µ
â å®åžžæ§, èªå·±çžé¢(ACF/PACF), ARIMAã¢ãã«, ç¶æ 空éã¢ãã«, ã«ã«ãã³ãã£ã«ã¿, ããŒã¹ã©ã€ã³ã¢ãã«ãšããŠã®éèŠæ§ - 第63åïŒC63ïŒéå»ãèšæ¶ããAIãRNNãšLSTM/GRUã®ã¡ã«ããºã
â ååž°çæ§é , é·æäŸåæ§, åŸé æ¶å€±ã»ççºåé¡, LSTM, GRU, ã²ãŒãæ©æ§ (å ¥å, å¿åŽ, åºå) - 第64åïŒC64ïŒæªæ¥ãäºæž¬ãã翻蚳æ©ãSeq2Seqãšã¢ãã³ã·ã§ã³æ©æ§
â ãšã³ã³ãŒãã»ãã³ãŒãã¢ãã«, Seq2Seq, æèãã¯ãã«, ã¢ãã³ã·ã§ã³æ©æ§, æç³»åäºæž¬, æ©æ¢°ç¿»èš³ã®ã¢ãããžãŒ - 第65åïŒC65ïŒç»åã ããããªãïŒæç³»åè§£æã«ãããCNNã®é ããå
â 1D-CNN, 屿çãã¿ãŒã³ã®æœåº, å æçç³ã¿èŸŒã¿, æ¡åŒµç³ã¿èŸŒã¿, å容é, WaveNet, TCN (Temporal Convolutional Networks) - 第66åïŒC66ïŒæç³»åè§£æã®é©åœå
ãTransformerã®åšåãšå¿çš
â Self-Attentionã«ããé·æäŸåæ§ã®ææ, Positional Encoding, èšç®å¹çã®èª²é¡, ææ°ã¢ãã« (Informer, PatchTST), å»çå¿çš- 第66.1åïŒC66.1ïŒãPost-TransformerãMamba & Sambaã§æãè¶
é·å°ºãã€ã¿ã«è§£æ
â State Space Models (SSM), Mamba architecture, Samba (Hybrid), èšç®å¹çãšç²ŸåºŠã®ãã¬ãŒããªã, mamba-ssm ã©ã€ãã©ãªå®è£
- 第66.1åïŒC66.1ïŒãPost-TransformerãMamba & Sambaã§æãè¶
é·å°ºãã€ã¿ã«è§£æ
- 第67åïŒC67ïŒè€æ°ã®æ
å ±ãçµ±åããããã«ãã¢ãŒãã«æç³»åè§£æ
â ãã«ãã¢ãŒãã«åŠç¿, éçããŒã¿ (æ£è èæ¯) ãšåçããŒã¿ (æç³»å) ã®çµ±å, ããŒã¿ãã¥ãŒãžã§ã³ (æ©æ/åŸæ), èšåºããã¹ããšã®é£æº - 第68åïŒC68ïŒAIã®æèããã»ã¹ãèŠããæç³»åã¢ãã«ã®XAIïŒèª¬æå¯èœãªAIïŒ
â 説æå¯èœãªAI (XAI), ã¢ãã³ã·ã§ã³ãããã®å¯èŠå, ç¹åŸŽééèŠåºŠ, LIME, SHAP, å€ææ ¹æ ã®ç¹å®, èšåºçä¿¡é Œæ§ - 第69åïŒC69ïŒãå®è·µãICUããŒã¿ã§æè¡çã®å
åãæããäºæž¬ã¢ãã«ãæ§ç¯ãã
â ICUããŒã¿ã»ãã (MIMIC-IV), ããŒã¿ååŠçãã€ãã©ã€ã³, ã¢ãã«æ§ç¯ãšèšç·Ž, æ§èœè©äŸ¡ (AUC, F1ã¹ã³ã¢), èšåºã·ããªãªã§ã®è©äŸ¡, çµæã®è§£é
第VIIéšïŒAIã”åµé ”ããæªæ¥ãçæã¢ãã«ã®äžç

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C70 | 第70åïŒå»çã«é©åœãïŒçæAIããããããã©ãã€ã ã·ãã | çæAI, åºç€ã¢ãã« (Foundation Models), å€§èŠæš¡èšèªã¢ãã«(LLM), ã¹ã±ãŒãªã³ã°å, åµçºçèœå, å»çå¿çšåé (èšºææ¯æŽ, åµè¬, åå¥åå»ç) |
| C71 | 第71åïŒååšããªãããŒã¿ãåµãåºããVAEãšGANã®é¬éè¡ | ãªãŒããšã³ã³ãŒã, æœåšç©ºé, å€åãªãŒããšã³ã³ãŒã(VAE), ELBO, æµå¯Ÿççæãããã¯ãŒã¯(GAN), çæåšãšèå¥åš, ããŒã¿æ¡åŒµ, åæããŒã¿ |
| C72 | 第72åïŒãã€ãºããçŸãçããæ¡æ£ã¢ãã«ã«ããè¶ ãªã¢ã«ç»åçæ | æ¡æ£ã¢ãã«(DDPM), é æ¹åéçšãšéæ¹åéçš, ã¹ã³ã¢ããŒã¹ã¢ãã«, ãã€ãºé€å», Stable Diffusion, èªå° (Guidance), åæå»çšç»å |
| C73 | 第73åïŒChatGPTã®é è³ãå€§èŠæš¡èšèªã¢ãã«(LLM)ã®ä»çµã¿ãè§£åãã | GPT, PaLM, Transformerãã³ãŒã, èªå·±æåž«ããåŠç¿, Next Token Prediction, ããã³ãããšã³ãžãã¢ãªã³ã°, In-Context Learning, Chain-of-Thought |
| C74 | 第74åïŒå·šå€§AIãå¹ççã«ã調æãããæè¡ãPEFTãšRLHF | ãã©ã¡ãŒã¿å¹ççãã¡ã€ã³ãã¥ãŒãã³ã°(PEFT), LoRA, QLoRA, 人éã®ãã£ãŒãããã¯ã«ãã匷ååŠç¿(RLHF), å ±é ¬ã¢ãã«, PPO |
| C75 | 第75åïŒç»åãšèšèãçè§£ããããã«ãã¢ãŒãã«AIã®è¡æ | ãã«ãã¢ãŒãã«åŠç¿, ç»åãšèšèªã®çµ±å (GPT-4V), CLIP (å¯Ÿç §åŠç¿), å ±ååã蟌ã¿ç©ºé, 蚺æã¬ããŒãèªåçæ, VQA (Visual Question Answering) |
| C76 | 第76åïŒAIç§åŠè ãèªçãèªåŸåãšãŒãžã§ã³ãã«ããç§åŠççºèŠã®èªåå | AIãšãŒãžã§ã³ã, ReAct (Reason+Act), ããŒã«å©çš, åµè¬, ã¿ã³ãã¯è³ªæ§é äºæž¬ (AlphaFold), ãªãµãŒããšãŒãžã§ã³ã, ä»®èª¬çæ |
| C77 | 第77åïŒä»®æ³æ£è ã§æªæ¥ãå ããããžã¿ã«ãã€ã³ãšIn Silicoèšåºè©Šéš | ããžã¿ã«ãã€ã³, ä»®æ³æ£è , çŸæ£é²è¡ã·ãã¥ã¬ãŒã·ã§ã³, æ²»çå¹æäºæž¬, In Silicoèšåºè©Šéš, åå¥åå»ç, äºæž¬å»ç |
| C78 | 第78åïŒAIããäžçããåŠã¶ãã¯ãŒã«ãã¢ãã«ãšèªå·±èªèãžã®é | ã¯ãŒã«ãã¢ãã«, ç°å¢ã®å éšã¢ãã«, æªæ¥äºæž¬, èšç» (Planning), æ³åå, 匷ååŠç¿ãšã®èå, AIã®èªåŸæ§ |
| C79 | 第79åïŒçæAIã®å ãšåœ±ã責任ããã€ãããŒã·ã§ã³ã®ããã®å«çãšã¬ããã³ã¹ | ãã«ã·ããŒã·ã§ã³, ãã€ã¢ã¹ãšå ¬å¹³æ§, ããŒã¿ãã©ã€ãã·ãŒ, èäœæš©ãšç¥ç財ç£, 説æè²¬ä»», åœæ å ±, å»çAIã®èŠå¶åå |
C70-79
- 第70åïŒC70ïŒå»çã«é©åœãïŒçæAIããããããã©ãã€ã ã·ãã
â çæAI, åºç€ã¢ãã« (Foundation Models), å€§èŠæš¡èšèªã¢ãã«(LLM), ã¹ã±ãŒãªã³ã°å, åµçºçèœå, å»çå¿çšåé (èšºææ¯æŽ, åµè¬, åå¥åå»ç) - 第71åïŒC71ïŒååšããªãããŒã¿ãåµãåºããVAEãšGANã®é¬éè¡
â ãªãŒããšã³ã³ãŒã, æœåšç©ºé, å€åãªãŒããšã³ã³ãŒã(VAE), ELBO, æµå¯Ÿççæãããã¯ãŒã¯(GAN), çæåšãšèå¥åš, ããŒã¿æ¡åŒµ, åæããŒã¿ - 第72åïŒC72ïŒãã€ãºããçŸãçããæ¡æ£ã¢ãã«ã«ããè¶
ãªã¢ã«ç»åçæ
â æ¡æ£ã¢ãã«(DDPM), é æ¹åéçšãšéæ¹åéçš, ã¹ã³ã¢ããŒã¹ã¢ãã«, ãã€ãºé€å», Stable Diffusion, èªå° (Guidance), åæå»çšç»å - 第73åïŒC73ïŒChatGPTã®é è³ãå€§èŠæš¡èšèªã¢ãã«(LLM)ã®ä»çµã¿ãè§£åãã
â GPT, PaLM, Transformerãã³ãŒã, èªå·±æåž«ããåŠç¿, Next Token Prediction, ããã³ãããšã³ãžãã¢ãªã³ã°, In-Context Learning, Chain-of-Thought - 第74åïŒC74ïŒå·šå€§AIãå¹ççã«ã調æãããæè¡ãPEFTãšRLHF
â ãã©ã¡ãŒã¿å¹ççãã¡ã€ã³ãã¥ãŒãã³ã°(PEFT), LoRA, QLoRA, 人éã®ãã£ãŒãããã¯ã«ãã匷ååŠç¿(RLHF), å ±é ¬ã¢ãã«, PPO - 第75åïŒC75ïŒç»åãšèšèãçè§£ããããã«ãã¢ãŒãã«AIã®è¡æ
â ãã«ãã¢ãŒãã«åŠç¿, ç»åãšèšèªã®çµ±å (GPT-4V), CLIP (å¯Ÿç §åŠç¿), å ±ååã蟌ã¿ç©ºé, 蚺æã¬ããŒãèªåçæ, VQA (Visual Question Answering) - 第76åïŒC76ïŒAIç§åŠè
ãèªçãèªåŸåãšãŒãžã§ã³ãã«ããç§åŠççºèŠã®èªåå
â AIãšãŒãžã§ã³ã, ReAct (Reason+Act), ããŒã«å©çš, åµè¬, ã¿ã³ãã¯è³ªæ§é äºæž¬ (AlphaFold), ãªãµãŒããšãŒãžã§ã³ã, ä»®èª¬çæ - 第77åïŒC77ïŒä»®æ³æ£è
ã§æªæ¥ãå ããããžã¿ã«ãã€ã³ãšIn Silicoèšåºè©Šéš
â ããžã¿ã«ãã€ã³, ä»®æ³æ£è , çŸæ£é²è¡ã·ãã¥ã¬ãŒã·ã§ã³, æ²»çå¹æäºæž¬, In Silicoèšåºè©Šéš, åå¥åå»ç, äºæž¬å»ç - 第78åïŒC78ïŒAIããäžçããåŠã¶ãã¯ãŒã«ãã¢ãã«ãšèªå·±èªèãžã®é
â ã¯ãŒã«ãã¢ãã«, ç°å¢ã®å éšã¢ãã«, æªæ¥äºæž¬, èšç» (Planning), æ³åå, 匷ååŠç¿ãšã®èå, AIã®èªåŸæ§ - 第79åïŒC79ïŒçæAIã®å
ãšåœ±ã責任ããã€ãããŒã·ã§ã³ã®ããã®å«çãšã¬ããã³ã¹
â ãã«ã·ããŒã·ã§ã³, ãã€ã¢ã¹ãšå ¬å¹³æ§, ããŒã¿ãã©ã€ãã·ãŒ, èäœæš©ãšç¥ç財ç£, 説æè²¬ä»», åœæ å ±, å»çAIã®èŠå¶åå
第VIIIéšïŒAIã”詊è¡é¯èª€”ã§åŠã¶ãæåŒ·ã®æææ±ºå®è¡

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C80 | 第80åïŒAIãèªãåŠã¶ïŒã匷ååŠç¿ããžã®æåŸ ç¶ | 匷ååŠç¿, ãšãŒãžã§ã³ã, ç°å¢, è¡å, å ±é ¬, æ¹ç(Policy), ãã«ã³ã決å®éçš(MDP), ç¶æ é·ç§»ç¢ºç, å²åŒç |
| C81 | 第81åïŒè¡åã®ã䟡å€ããæž¬ãçŸ éç€ããã«ãã³æ¹çšåŒãšäŸ¡å€ããŒã¹åŠç¿ | 䟡å€é¢æ° (ç¶æ 䟡å€/è¡å䟡å€), ãã«ãã³æ¹çšåŒ (æåŸ /æé©), åçèšç»æ³, 䟡å€å埩æ³, æ¹çååŸ©æ³ |
| C82 | 第82åïŒäŸ¡å€ããè¡åãéžã¶AIãQåŠç¿ãšDQNïŒæ·±å±€Qãããã¯ãŒã¯ïŒ | QåŠç¿, QããŒãã«, æéå·®åŠç¿(TDåŠç¿), SARSA, DQN, çµéšåç (Experience Replay), ã¿ãŒã²ãããããã¯ãŒã¯, Double DQN |
| C83 | 第83åïŒãæ¹éããçŽæ¥ç£šãäžããAIãæ¹çåŸé æ³ãšActor-Critic | æ¹çåŸé æ³, REINFORCE, ããŒã¹ã©ã€ã³, Actor-Critic (A2C, A3C), AdvantageïŒåªäœé¢æ°ïŒ |
| C84 | 第84åïŒå®å®ãšå¹çã远æ±ãããçŸä»£çã¢ã¯ã¿ïœ°ã¯ãªãã£ãã¯ææ³ | PPO (Proximal Policy Optimization), TRPO, é£ç¶å€å¶åŸ¡, SAC (Soft Actor-Critic), æå€§ãšã³ããããŒåŒ·ååŠç¿ |
| C85 | 第85åïŒæªç¥ãžã®ææŠããæåæã®æŽ»çšãïŒãæ¢çŽ¢ãšæŽ»çšãã®ãžã¬ã³ããšæŠç¥ | ε-greedyæ³, UCB (Upper Confidence Bound), ãã³ããœã³ãµã³ããªã³ã°, 奜å¥å¿é§ååæ¢çŽ¢, å çºçåæ©ã¥ã |
| C86 | 第86åïŒéå»ã®å»çããŒã¿ãããæé©ããåŠã¶ããªãã©ã€ã³åŒ·ååŠç¿ | ãªãã©ã€ã³åŒ·ååŠç¿, ããã匷ååŠç¿, ååžå€(OOD)åé¡, Conservative Q-Learning (CQL), ãªãããªã·ãŒè©äŸ¡ (OPE) |
| C87 | 第87åïŒAIã®äžã«ãä»®æ³æ£è ããåµããã¢ãã«ããŒã¹åŒ·ååŠç¿ | ã¢ãã«ããŒã¹åŒ·ååŠç¿, ç°å¢ã¢ãã«(ã¯ãŒã«ãã¢ãã«), Dyna-Q, ã·ãã¥ã¬ãŒã·ã§ã³ã«ããèšç»(Planning), ããŒã¿å¹ç |
| C88 | 第88åïŒã決ããŠè¶ ããŠã¯ãªããªãäžç·ããæãããå®å šãªåŒ·ååŠç¿ãšã¢ã©ã€ã¡ã³ã | å®å šãªåŒ·ååŠç¿, å¶çŽä»ãMDP(CMDP), 人éã®ãã£ãŒãããã¯ã«ãã匷ååŠç¿(RLHF), å ±é ¬ã¢ããªã³ã°, ã¢ã©ã€ã¡ã³ã, è§£éå¯èœæ§(XAI for RL) |
| C89 | 第89åïŒãå¿çšãèšåºã®æææ±ºå®ãæ¯æŽãããåçæ²»çèšç»ãšå®è£ ãžã®é | åçæ²»çèšç»(DTR), åå¥åå»ç, èšåºå¿çš (æè¡ç管ç, ããæ²»ç), ã·ãã¥ã¬ãŒã·ã§ã³ç°å¢, èŠå¶ãšå«ççèª²é¡ |
C80-89
- 第80åïŒC80ïŒAIãèªãåŠã¶ïŒã匷ååŠç¿ããžã®æåŸ
ç¶
â 匷ååŠç¿, ãšãŒãžã§ã³ã, ç°å¢, è¡å, å ±é ¬, æ¹ç(Policy), ãã«ã³ã決å®éçš(MDP), ç¶æ é·ç§»ç¢ºç, å²åŒç - 第81åïŒC81ïŒè¡åã®ã䟡å€ããæž¬ãçŸ
éç€ããã«ãã³æ¹çšåŒãšäŸ¡å€ããŒã¹åŠç¿
â 䟡å€é¢æ° (ç¶æ 䟡å€/è¡å䟡å€), ãã«ãã³æ¹çšåŒ (æåŸ /æé©), åçèšç»æ³, 䟡å€å埩æ³, æ¹çååŸ©æ³ - 第82åïŒC82ïŒäŸ¡å€ããè¡åãéžã¶AIãQåŠç¿ãšDQNïŒæ·±å±€Qãããã¯ãŒã¯ïŒ
â QåŠç¿, QããŒãã«, æéå·®åŠç¿(TDåŠç¿), SARSA, DQN, çµéšåç (Experience Replay), ã¿ãŒã²ãããããã¯ãŒã¯, Double DQN - 第83åïŒC83ïŒãæ¹éããçŽæ¥ç£šãäžããAIãæ¹çåŸé
æ³ãšActor-Critic
â æ¹çåŸé æ³, REINFORCE, ããŒã¹ã©ã€ã³, Actor-Critic (A2C, A3C), AdvantageïŒåªäœé¢æ°ïŒ - 第84åïŒC84ïŒå®å®ãšå¹çã远æ±ãããçŸä»£çã¢ã¯ã¿ïœ°ã¯ãªãã£ãã¯ææ³
â PPO (Proximal Policy Optimization), TRPO, é£ç¶å€å¶åŸ¡, SAC (Soft Actor-Critic), æå€§ãšã³ããããŒåŒ·ååŠç¿ - 第85åïŒC85ïŒæªç¥ãžã®ææŠããæåæã®æŽ»çšãïŒãæ¢çŽ¢ãšæŽ»çšãã®ãžã¬ã³ããšæŠç¥
â ε-greedyæ³, UCB (Upper Confidence Bound), ãã³ããœã³ãµã³ããªã³ã°, 奜å¥å¿é§ååæ¢çŽ¢, å çºçåæ©ã¥ã - 第86åïŒC86ïŒéå»ã®å»çããŒã¿ãããæé©ããåŠã¶ããªãã©ã€ã³åŒ·ååŠç¿
â ãªãã©ã€ã³åŒ·ååŠç¿, ããã匷ååŠç¿, ååžå€(OOD)åé¡, Conservative Q-Learning (CQL), ãªãããªã·ãŒè©äŸ¡ (OPE) - 第87åïŒC87ïŒAIã®äžã«ãä»®æ³æ£è
ããåµããã¢ãã«ããŒã¹åŒ·ååŠç¿
â ã¢ãã«ããŒã¹åŒ·ååŠç¿, ç°å¢ã¢ãã«(ã¯ãŒã«ãã¢ãã«), Dyna-Q, ã·ãã¥ã¬ãŒã·ã§ã³ã«ããèšç»(Planning), ããŒã¿å¹ç - 第88åïŒC88ïŒã決ããŠè¶
ããŠã¯ãªããªãäžç·ããæãããå®å
šãªåŒ·ååŠç¿ãšã¢ã©ã€ã¡ã³ã
â å®å šãªåŒ·ååŠç¿, å¶çŽä»ãMDP(CMDP), 人éã®ãã£ãŒãããã¯ã«ãã匷ååŠç¿(RLHF), å ±é ¬ã¢ããªã³ã°, ã¢ã©ã€ã¡ã³ã, è§£éå¯èœæ§(XAI for RL) - 第89åïŒC89ïŒãå¿çšãèšåºã®æææ±ºå®ãæ¯æŽãããåçæ²»çèšç»ãšå®è£
ãžã®é
â åçæ²»çèšç»(DTR), åå¥åå»ç, èšåºå¿çš (æè¡ç管ç, ããæ²»ç), ã·ãã¥ã¬ãŒã·ã§ã³ç°å¢, èŠå¶ãšå«çç課é¡
第IXéšïŒç 究宀ããäžçãžãå»çAI瀟äŒå®è£ ããŒãããã

| ID | ã¿ã€ãã« | æŠèŠã»ããŒã¯ãŒã |
|---|---|---|
| C90 | 第90åïŒã¢ã€ãã¢ãã補åãžãå»çAIãããžã§ã¯ãã®ã©ã€ããµã€ã¯ã«å šè² | 課é¡å®çŸ© (èšåºããŒãº), PoC (æŠå¿µå®èšŒ), MVP (å®çšæå°éã®è£œå), ã¢ãžã£ã€ã«éçº, V&V (æ€èšŒãšåŠ¥åœæ§ç¢ºèª), è¬äºç³è«, äžåž, éçšã»ä¿å® (Post-Market Surveillance) |
| C91 | 第91åïŒæåŒ·ããŒã ãäœãïŒã¢ãžã£ã€ã«éçºãšã¢ãã³ãªããŒã ãã«ãã£ã³ã° | ã¹ã¯ã©ã (ã¹ããªã³ã, ãã€ãªãŒã¹ã¿ã³ãã¢ãã), ã«ã³ãã³, 圹å²åæ (å»åž«, ãšã³ãžãã¢, PM, ãã¶ã€ããŒ), RACIãã£ãŒã, ã¹ããŒã¯ãã«ããŒç®¡ç, ã³ãã¥ãã±ãŒã·ã§ã³èšèš |
| C92 | 第92åïŒ1幎åŸãåãã³ãŒããæžããä¿¡é Œæ§ã®ããã®ãœãããŠã§ã¢å·¥åŠ | ã³ãŒãå質 (PEP8, Black, isort, flake8), ããŒãžã§ã³ç®¡ç (Git, GitHub, ãã©ã³ãæŠç¥), ãã¹ã (åäœãã¹ã/pytest, çµåãã¹ã), ããã¥ã¡ã³ããŒã·ã§ã³ (docstring, Sphinx), ã³ãŒãã¬ãã¥ãŒ |
| C93 | 第93åïŒéçºãèªååã»å¹çåããMLOpsïŒæ©æ¢°åŠç¿åºç€ïŒã®éæ³ | CI/CD (ç¶ç¶çã€ã³ãã°ã¬ãŒã·ã§ã³/ããªããªãŒ), ã³ã³ããå (Docker), ãã€ãã©ã€ã³ãªãŒã±ã¹ãã¬ãŒã·ã§ã³ (MLflow, Kubeflow), ã¢ãã«ã¬ãžã¹ããª, ã¢ãã«ç£èŠ (ããŒã¿ããªãã, ã³ã³ã»ããããªãã) |
| C94 | 第94åïŒç é¢ã·ã¹ãã ãšAIãç¹ããå»çæ å ±é£æºè¡ | é»åã«ã«ã (EHR), PACS, å»çæšæºèŠæ Œ (HL7v2, FHIR, DICOM), IHE, é¢å ãããã¯ãŒã¯, VPN, ããŒã¿é£æºãšçžäºéçšæ§ |
| C95 | 第95åïŒæ£è æ å ±ãå®ãæããã»ãã¥ã¢ãªã¯ã©ãŠãã€ã³ãã©èšèš | ã¯ã©ãŠã (AWS/GCP/Azure), HIPAA, 3ç2ã¬ã€ãã©ã€ã³, ISMAP, ã»ãã¥ã¢ãªã¢ãŒããã¯ã㣠(VPC, IAM), æå·å (éä¿¡/ä¿ç®¡), ã¢ã¯ã»ã¹å¶åŸ¡, ç£æ»ãã° |
| C96 | 第96åïŒãå»çæ©åšããšããŠã®AIãèŠå¶ãšè¬äºæŠç¥ãä¹ãè¶ãã | å»çæ©åšããã°ã©ã (SaMD), è¬æ©æ³, FDA/CEããŒãã³ã°, å質ãããžã¡ã³ãã·ã¹ãã (QMS, ISO13485), ãœãããŠã§ã¢ã©ã€ããµã€ã¯ã« (IEC62304), ãªã¹ã¯ãããžã¡ã³ã (ISO14971), èšåºè©äŸ¡ |
| C97 | 第97åïŒããã€ã¹ã®äžã§åããããšããžAIãšIoTå»çæ©åš | ãšããžAI, çµã¿èŸŒã¿ã·ã¹ãã , ãªã¢ã«ã¿ã€ã æšè«, ã¢ãã«è»œéå (éåå, èžç), ãŠã§ã¢ã©ãã«ããã€ã¹, IoT (Internet of Things), ã»ã³ãµãŒé£æº (BLE) |
| C98 | 第98åïŒå»çAIããžãã¹ãåµãïŒäºæ¥èšç»ãšç¥è²¡æŠç¥ | ããžãã¹ã¢ãã«ãã£ã³ãã¹, äºæ¥èšç», ããã¿ã€ãº (課éã¢ãã«), ããŒã±ããåæ, ç¥çè²¡ç£ (ç¹èš±, èäœæš©, ããŠããŠ), è«æçºè¡šæŠç¥ |
| C99 | 第99åïŒã忥å¶äœãæªæ¥ã®å»çAIãµãŒãã¹ãäŒç»ã»èšèšãã | ç·åæŒç¿, èšåºèª²é¡èšå®, ããžãã¹ãã©ã³äœæ, æè¡ã¢ãŒããã¯ãã£èšèš, èŠå¶å¯Ÿå¿èšç», ãããã¿ã€ãéçº, æè³å®¶åããã¬ãŒã³ããŒã·ã§ã³ |
C90-99
- 第90åïŒC90ïŒã¢ã€ãã¢ãã補åãžãå»çAIãããžã§ã¯ãã®ã©ã€ããµã€ã¯ã«å
šè²
â 課é¡å®çŸ© (èšåºããŒãº), PoC (æŠå¿µå®èšŒ), MVP (å®çšæå°éã®è£œå), ã¢ãžã£ã€ã«éçº, V&V (æ€èšŒãšåŠ¥åœæ§ç¢ºèª), è¬äºç³è«, äžåž, éçšã»ä¿å® (Post-Market Surveillance) - 第91åïŒC91ïŒæåŒ·ããŒã ãäœãïŒã¢ãžã£ã€ã«éçºãšã¢ãã³ãªããŒã ãã«ãã£ã³ã°
â ã¹ã¯ã©ã (ã¹ããªã³ã, ãã€ãªãŒã¹ã¿ã³ãã¢ãã), ã«ã³ãã³, 圹å²åæ (å»åž«, ãšã³ãžãã¢, PM, ãã¶ã€ããŒ), RACIãã£ãŒã, ã¹ããŒã¯ãã«ããŒç®¡ç, ã³ãã¥ãã±ãŒã·ã§ã³èšèš - 第92åïŒC92ïŒ1幎åŸãåãã³ãŒããæžããä¿¡é Œæ§ã®ããã®ãœãããŠã§ã¢å·¥åŠ
â ã³ãŒãå質 (PEP8, Black, isort, flake8), ããŒãžã§ã³ç®¡ç (Git, GitHub, ãã©ã³ãæŠç¥), ãã¹ã (åäœãã¹ã/pytest, çµåãã¹ã), ããã¥ã¡ã³ããŒã·ã§ã³ (docstring, Sphinx), ã³ãŒãã¬ãã¥ãŒ - 第93åïŒC93ïŒéçºãèªååã»å¹çåããMLOpsïŒæ©æ¢°åŠç¿åºç€ïŒã®éæ³
â CI/CD (ç¶ç¶çã€ã³ãã°ã¬ãŒã·ã§ã³/ããªããªãŒ), ã³ã³ããå (Docker), ãã€ãã©ã€ã³ãªãŒã±ã¹ãã¬ãŒã·ã§ã³ (MLflow, Kubeflow), ã¢ãã«ã¬ãžã¹ããª, ã¢ãã«ç£èŠ (ããŒã¿ããªãã, ã³ã³ã»ããããªãã) - 第94åïŒC94ïŒç
é¢ã·ã¹ãã ãšAIãç¹ããå»çæ
å ±é£æºè¡
â é»åã«ã«ã (EHR), PACS, å»çæšæºèŠæ Œ (HL7v2, FHIR, DICOM), IHE, é¢å ãããã¯ãŒã¯, VPN, ããŒã¿é£æºãšçžäºéçšæ§ - 第95åïŒC95ïŒæ£è
æ
å ±ãå®ãæããã»ãã¥ã¢ãªã¯ã©ãŠãã€ã³ãã©èšèš
â ã¯ã©ãŠã (AWS/GCP/Azure), HIPAA, 3ç2ã¬ã€ãã©ã€ã³, ISMAP, ã»ãã¥ã¢ãªã¢ãŒããã¯ã㣠(VPC, IAM), æå·å (éä¿¡/ä¿ç®¡), ã¢ã¯ã»ã¹å¶åŸ¡, ç£æ»ãã° - 第96åïŒC96ïŒãå»çæ©åšããšããŠã®AIãèŠå¶ãšè¬äºæŠç¥ãä¹ãè¶ãã
â å»çæ©åšããã°ã©ã (SaMD), è¬æ©æ³, FDA/CEããŒãã³ã°, å質ãããžã¡ã³ãã·ã¹ãã (QMS, ISO13485), ãœãããŠã§ã¢ã©ã€ããµã€ã¯ã« (IEC62304), ãªã¹ã¯ãããžã¡ã³ã (ISO14971), èšåºè©äŸ¡ - 第97åïŒC97ïŒããã€ã¹ã®äžã§åããããšããžAIãšIoTå»çæ©åš
â ãšããžAI, çµã¿èŸŒã¿ã·ã¹ãã , ãªã¢ã«ã¿ã€ã æšè«, ã¢ãã«è»œéå (éåå, èžç), ãŠã§ã¢ã©ãã«ããã€ã¹, IoT (Internet of Things), ã»ã³ãµãŒé£æº (BLE) - 第98åïŒC98ïŒå»çAIããžãã¹ãåµãïŒäºæ¥èšç»ãšç¥è²¡æŠç¥
â ããžãã¹ã¢ãã«ãã£ã³ãã¹, äºæ¥èšç», ããã¿ã€ãº (課éã¢ãã«), ããŒã±ããåæ, ç¥çè²¡ç£ (ç¹èš±, èäœæš©, ããŠããŠ), è«æçºè¡šæŠç¥ - 第99åïŒC99ïŒã忥å¶äœãæªæ¥ã®å»çAIãµãŒãã¹ãäŒç»ã»èšèšãã
â ç·åæŒç¿, èšåºèª²é¡èšå®, ããžãã¹ãã©ã³äœæ, æè¡ã¢ãŒããã¯ãã£èšèš, èŠå¶å¯Ÿå¿èšç», ãããã¿ã€ãéçº, æè³å®¶åããã¬ãŒã³ããŒã·ã§ã³
ãå©çšèŠçŽïŒå 責äºé ïŒ
åœãµã€ãïŒä»¥äžãæ¬ãµã€ãããšãããŸãïŒããå©çšã«ãªãåã«ãæ¬ãå©çšèŠçŽïŒä»¥äžãæ¬èŠçŽããšãããŸãïŒããããèªã¿ãã ãããæ¬ãµã€ããå©çšãããæç¹ã§ãå©çšè ã¯æ¬èŠçŽã®å šãŠã®æ¡é ã«åæãããã®ãšã¿ãªããŸãã
第1æ¡ïŒç®çãšæ å ±ã®æ§è³ªïŒ
- æ¬ãµã€ãã¯ãå»çåéã«ãããAIæè¡ã«é¢ããäžè¬çãªæ å ±æäŸããã³æè¡çãªåŠç¿æ©äŒã®æäŸãå¯äžã®ç®çãšããŸãã
- æ¬ãµã€ãã§æäŸããããã¹ãŠã®ã³ã³ãã³ãïŒæç« ãå³è¡šãã³ãŒããããŒã¿ã»ããã®ç޹ä»çãå«ã¿ãŸããããããã«éå®ãããŸããïŒã¯ãäžè¬çãªåŠç¿åèçšã§ããããããªãå Žåãå»åŠçãªå©èšã蚺æãæ²»çããŸãã¯ãããã«æºããè¡çºïŒä»¥äžãå»è¡çºçããšãããŸãïŒãæäŸãããã®ã§ã¯ãããŸããã
- æ¬ãµã€ãã®ã³ã³ãã³ãã¯ãç¹å®ã®è£œåãæè¡ããŸãã¯æ²»çæ³ã®æå¹æ§ãå®å šæ§ãä¿èšŒãæšå¥šããŸãã¯åºåã»è²©å£²ä¿é²ãããã®ã§ã¯ãããŸããã玹ä»ããæè¡ã«ã¯ç ç©¶éçºæ®µéã®ãã®ãå«ãŸããŠããããã®èšåºå¿çšã«ã¯ããããªãç ç©¶ãšåœå å€ã®èŠå¶åœå±ã«ããæ£åŒãªæ¿èªãå¥éå¿ èŠã§ãã
- æ¬ãµã€ãã¯ãæ å ±æäŸãç®çãšãããã®ã§ãããç¹å®ã®æ²»çæ³ãæšå¥šãããã®ã§ã¯ãããŸãããå¥åº·ã«é¢ãããæžå¿µããçžè«ã¯ãå¿ ãå°éã®å»çæ©é¢ã«ãçžè«ãã ããã
第2æ¡ïŒæ³ä»€çã®éµå®ïŒ
å©çšè
ã¯ãæ¬ãµã€ãã®å©çšã«ããããå»åž«æ³ãå»è¬åãå»çæ©åšçã®å質ãæå¹æ§åã³å®å
šæ§ã®ç¢ºä¿çã«é¢ããæ³åŸïŒè¬æ©æ³ïŒãå人æ
å ±ã®ä¿è·ã«é¢ããæ³åŸãå»çæ³ãå»çåºåã¬ã€ãã©ã€ã³ããã®ä»é¢é£ããåœå
å€ã®å
šãŠã®æ³ä»€ãæ¡äŸãèŠåãããã³åçåºã»åŠäŒçãå®ããææ°ã®ã¬ã€ãã©ã€ã³çããèªãã®è²¬ä»»ã«ãããŠéµå®ãããã®ãšããŸãããããã®é©çšå€æã«ã€ããŠããå©çšè
ãèªãé¢ä¿åæã«ç¢ºèªãããã®ãšããæ¬ãµã€ãã¯äžåã®è²¬ä»»ãè² ããŸããã
第3æ¡ïŒå»çè¡çºã«ããã責任ïŒ
- æ¬ãµã€ãã§ç޹ä»ããAIæè¡ã»ææ³ã¯ããããŸã§ç 究段éã®æè¡ç解説ã§ãããå®éã®èšåºçŸå Žã§ã®èšºæã»æ²»çã代æ¿ãè£å©ããŸãã¯æšå¥šãããã®ã§ã¯äžåãããŸããã
- å»è¡çºçã«é¢ããæçµçãªå€æã決å®ãããã³ããã«äŒŽãäžåã®è²¬ä»»ã¯ãå¿ ãæ³åŸäžãã®è³æ Œãèªããããå»çå°éå®¶ïŒå»åž«ãæ¯ç§å»åž«çïŒãè² ããã®ãšããŸããAIã«ããåºåããè³æ Œãæããå°éå®¶ã«ããç¬ç«ããæ€èšŒããã³å€æãçµãã«å©çšããããšãåºãçŠããŸãã
- æ¬ãµã€ãã®æ å ±ã«åºã¥ããããªãè¡çºã«ãã£ãŠå©çšè ãŸãã¯ç¬¬äžè ã«æå®³ãçããå Žåããæ¬ãµã€ãéå¶è ã¯äžåã®è²¬ä»»ãè² ããŸãããå®éã®èšåºå€æã«éããŠã¯ãå¿ ãæ åœã®å»çå°éå®¶ã«ãçžè«ãã ãããæ¬ãµã€ãã®å©çšã«ãã£ãŠãå©çšè ãšæ¬ãµã€ãéå¶è ã®éã«ãå»åž«ãšæ£è ã®é¢ä¿ããŸãã¯ãã®ä»ãããªãå°éçãªé¢ä¿ãæç«ãããã®ã§ã¯ãããŸããã
第4æ¡ïŒæ å ±ã®æ£ç¢ºæ§ã»å®å šæ§ã»æçšæ§ïŒ
- æ¬ãµã€ãã¯ãæ²èŒããæ å ±ïŒæ°å€ãäºäŸããœãŒã¹ã³ãŒããã©ã€ãã©ãªã®ããŒãžã§ã³çïŒã®æ£ç¢ºæ§ãå®å šæ§ãç¶²çŸ æ§ãæçšæ§ãç¹å®ç®çãžã®é©åæ§ããã®ä»äžåã®äºé ã«ã€ããŠãäœãä¿èšŒãããã®ã§ã¯ãããŸããã
- æ²èŒæ å ±ã¯å·çæç¹ã®ãã®ã§ãããäºåãªã倿ŽãŸãã¯åé€ãããããšããããŸãããŸããæè¡ã®é²å±ãã©ã€ãã©ãªã®æŽæ°çã«ãããæ å ±ã¯å€ããªãå¯èœæ§ããããŸããå©çšè ã¯ãå¿ ãèªèº«ã§å ¬åŒããã¥ã¡ã³ãçã®ææ°æ å ±ã確èªããèªãã®è²¬ä»»ã§æ å ±ãå©çšãããã®ãšããŸãã
第5æ¡ïŒAIçæã³ã³ãã³ãã«é¢ããæ³šæäºé
ïŒ
æ¬ãµã€ãã®ã³ã³ãã³ãã«ã¯ãAIã«ããææ¡ãåºã«äœæãããéšåãå«ãŸããå ŽåããããŸãããå
¬éã«ããã£ãŠã¯äººéã«ããç£ä¿®ã»ç·šéãçµãŠããŸããå©çšè
ãçæAIçãçšããéã¯ããã«ã·ããŒã·ã§ã³ïŒäºå®ã«åºã¥ããªãæ
å ±ã®çæïŒããã€ã¢ã¹ã®ãªã¹ã¯ãå
åšããããšãååã«çè§£ãããã®åºåãéµåã¿ã«ããããšãªããå¿
ãå°éå®¶ã«ããæ€èšŒãè¡ããã®ãšããŸãã
第6æ¡ïŒç¥çè²¡ç£æš©ïŒ
- æ¬ãµã€ããæ§æãããã¹ãŠã®ã³ã³ãã³ãã«é¢ããèäœæš©ãåæšæš©ããã®ä»äžåã®ç¥çè²¡ç£æš©ã¯ãæ¬ãµã€ãéå¶è ãŸãã¯æ£åœãªæš©å©ãæãã第äžè ã«åž°å±ããŸãã
- æ¬ãµã€ãã®ã³ã³ãã³ããåŒçšã転èŒãè€è£œãæ¹å€ããã®ä»ã®äºæ¬¡å©çšãè¡ãå Žåã¯ãèäœæš©æ³ãã®ä»é¢é£æ³èŠãéµå®ããå¿ ãåºå žãæèšãããšãšãã«ãæš©å©è ã®èš±è«ŸãåŸããªã©ãé©åãªæç¶ããèªãã®è²¬ä»»ã§è¡ããã®ãšããŸãã
第7æ¡ïŒãã©ã€ãã·ãŒã»å«çïŒ
æ¬ãµã€ãã§ç޹ä»ãŸãã¯èšåãããããŒã¿ã»ããçãå©çšããå Žåãå©çšè
ã¯åœè©²ããŒã¿ã»ããã«ä»éããã©ã€ã»ã³ã¹æ¡ä»¶ããã³ç ç©¶å«çæéã峿 Œã«éµå®ããå人æ
å ±ã®å¿ååãåæååŸã®ç¢ºèªãªã©ãé©çšãããæ³èŠå¶ã«åºã¥ãå¿
èŠãšããããã¹ãŠã®æªçœ®ããèªãã®è²¬ä»»ã«ãããŠè¬ãããã®ãšããŸãã
第8æ¡ïŒå©çšç°å¢ïŒ
æ¬ãµã€ãã§ç޹ä»ãããœãŒã¹ã³ãŒããã©ã€ãã©ãªã¯ãå·çæç¹ã§ç¹å®ã®ããŒãžã§ã³ããã³å®è¡ç°å¢ïŒOSãããŒããŠã§ã¢ãäŸåããã±ãŒãžçïŒãåæãšããŠããŸããå©çšè
ã®ç°å¢ã«ãããåäœãä¿èšŒãããã®ã§ã¯ãªããäºææ§ã®åé¡çã«èµ·å ãããããªãäžå©çã»æå®³ã«ã€ããŠããæ¬ãµã€ãéå¶è
ã¯è²¬ä»»ãè² ããŸããã
第9æ¡ïŒå 責äºé ïŒ
- æ¬ãµã€ãéå¶è ã¯ãå©çšè ãæ¬ãµã€ããå©çšããããšããŸãã¯å©çšã§ããªãã£ãããšã«ãã£ãŠçããäžåã®æå®³ïŒçŽæ¥æå®³ã鿥æå®³ãä»éçæå®³ãç¹å¥æå®³ãæ²çœ°çæå®³ãéžå€±å©çãããŒã¿ã®æ¶å€±ãããã°ã©ã ã®æ¯æçãå«ã¿ãŸããããããã«éå®ãããŸããïŒã«ã€ããŠããã®åå ã®åŠäœãåãããäžåã®æ³ç責任ãè² ããªããã®ãšããŸãã
- æ¬ãµã€ãã®å©çšã¯ãåŠç¿ããã³ç ç©¶ç®çã«éå®ããããã®ãšãããã以å€ã®ç®çã§ã®å©çšã¯ãé æ ®ãã ããã
- æ¬ãµã€ãã®å©çšã«é¢é£ããŠãå©çšè ãšç¬¬äžè ãšã®éã§çŽäºãçããå Žåãå©çšè ã¯èªãã®è²»çšãšè²¬ä»»ã«ãããŠããã解決ãããã®ãšããæ¬ãµã€ãéå¶è ã«äžåã®è¿·æãŸãã¯æå®³ãäžããªããã®ãšããŸãã
- æ¬ãµã€ãéå¶è ã¯ããã€ã§ãäºåãªãæ¬ãµã€ãã®éå¶ãäžæãäžæ¢ããŸãã¯å 容ã倿Žã§ãããã®ãšããããã«ãã£ãŠå©çšè ã«çãããããªãæå®³ã«ã€ããŠã責任ãè² ããŸããã
第10æ¡ïŒèŠçŽã®å€æŽïŒ
æ¬ãµã€ãéå¶è
ã¯ãå¿
èŠãšå€æããå Žåãå©çšè
ã®æ¿è«ŸãåŸãããšãªãããã€ã§ãæ¬èŠçŽã倿Žããããšãã§ããŸãã倿ŽåŸã®èŠçŽã¯ãæ¬ãµã€ãäžã«æ²èŒãããæç¹ã§å¹åãçãããã®ãšããå©çšè
ã¯å€æŽåŸã®èŠçŽã«ææããããã®ãšããŸãã
第11æ¡ïŒæºæ æ³ããã³åæç®¡èœïŒ
æ¬èŠçŽã®è§£éã«ããã£ãŠã¯ãæ¥æ¬æ³ãæºæ æ³ãšããŸããæ¬ãµã€ãã®å©çšããã³æ¬èŠçŽã«é¢é£ããŠçããäžåã®çŽäºã«ã€ããŠã¯ãæ±äº¬å°æ¹è£å€æã第äžå¯©ã®å°å±çåæç®¡èœè£å€æãšããŸãã
For J³, may joy follow you.

