ãã£ã¼ãã©ã¼ãã³ã°ã¯ç¿ãããæ £ãããã
ããã¡ãã·ã§ã³ã§ãã£ã¼ãã©ã¼ãã³ã°ããã¦ãããã£ã¼ãã©ã¼ãã³ã°è¸äººããããã§ããç§ã¯ãç¹ã«å°é家ã§ãä½ã§ããªãã®ã§ãããæ©æ¢°å¦ç¿ã«èå³ãã£ã¦ããã£ã¼ãã©ã¼ãã³ã°ã«é¢ãããã¨ããã°ã§ã¢ã¦ãããããã¦ãããã¡ã«ãAIã«é¢ããæ¬ã«ååãã¯ã¬ã¸ããããããããAI解æã®ãªã³ã©ã¤ã³ã³ã³ãã¹ã#Aidemynoteã§ç¹å¥è³åè³ãããããã©ãºãã¤ãã¬ã¸ã³ã¨ããåæ¥èªã«AIé¢ä¿ã§è¨äºãæ¸ããããããã£ã¼ãã©ã¼ãã³ã°ããããã®è¨äºãããºã£ããããITmediaã§åãä¸ããããããã¨å¤å°ãªãã¨ã価å¤ãæä¾ã§ããããã«ãªã£ã¦ãã¾ããã
ãä½ã®ç¥èãããã¯ã°ã©ã¦ã³ããã大ããè½åãç¡ãèªåããã©ããã£ã¦ç¥èã身ã«ã¤ãããã¨ãã§ãããã¨ããã¨ãè²ã æ¬ãèªãã ã®ã§ãããå®éã«æãåããã¦ããã°ã©ã ãçµãã§ãå®åé¡ã«å¯¾ãã¦è©¦è¡é¯èª¤ããçµæãããã°ã«ã¢ã¦ãããããç¶ãããã¨ã大ããã®ããªã¨æãã¾ãã
ã家ã«ç©ã¾ããæ¬ã®1é¨ã以ä¸ãããã¾ãè¯ãç¡ãã£ãæ¬ãé»åæ¸ç±ããã®æ°åããã®ã§ãåè¨ã¯50åãããã«ãªããã
ãæ£ç´ãæªã ã«æ°å¼ã¨ãã«é¢ãã¦ã¯ããã£ã±ãåãããã¡ããªã®ã§ãããå人çã«ã¯ãã£ã¼ãã©ã¼ãã³ã°ã¯ç¿ãããæ £ãããéè¦ããªã¨æãã¦ãã¾ããã¾ãã¯æãåããã¦ãè²ã 試ãã¦çåã«æã£ãã¨ãããæ¬ããããã®æ å ±ãè«æãªã©ãèªãã§è£å®ãããã¨ããé²ãæ¹ãä¸çªèªåã«ã¨ã£ã¦ã¯å¹çãããã£ãããªã¨æã£ã¦ãã¾ãã
ãã£ã¼ãã©ã¼ãã³ã°ã®å¦ç¿ã«å¯¾ãã沢山ã®ã¯ã
ãæãåããã®ãéè¦ã¨ã¯è¨ãã¾ãããããã£ã±ãå®éã«ä½ããä½ãã®ã¯å¤§å¤ã§ãããã£ã¼ãã©ã¼ãã³ã°ã®å¦ç¿ã«ã¯è²ã ãªç½ ãããã¾ãã
- ã»ããã¢ãããå°çã§ãç°å¢è¨å®ãã§ããé ã«ã¯åå°½ãã¦ããç½
- TensorFlowãKerasãChainerãPyTorchã¨ãããããã¬ã¼ã ã¯ã¼ã¯ããã£ã¦ã©ããè¯ãã®ããã£ã±ãããããªãç½
- ç°å¢è¨å®ãã§ããã¨ããã§ããã¼ã¸ã§ã³ã®éãã§ãµã³ãã«ãåããªãç½
- 大éã«æ¬ããã£ã¦ãã©ããè¯ãã®ãããåãããªãç½ ãè¼ã£ã¦ããããã°ã©ã åããæ¹ãåãããªãç½
- ãµã³ãã«ãã¯ã½ã¤ã¾ããªãï¼çµæã®æå³ãä¸æãªç½ ï¼èªåã¯ãæåMNISTãæåèªèã®ã¿ã¹ã¯ã¨ãããã¨ããåãã£ã¦ãªãã£ãï¼
- ãµã³ãã«ããå ã«ä¸åé²ããªããä½ããããè¯ããããããªãç½
ãã©ããä¸ã¤ã§ãå¿å½ããã®ãã人ã¯å¤ãã®ã§ã¯ç¡ãã§ããããï¼ç§ã¯ãå ¨é¨ã®ã¯ãã«å¼ã£ããã£ã¦ããã«ã»ã«ã¯ã®èã§ã®ã¬ããã®å¦ããè²ã ç ç²ã«ããªããé§ãæãã¦ãã¾ããï¼ä¸»ã«ç¡ç æéã¨ãéï¼
ãã£ã¼ãã©ã¼ãã³ã°ã®å¦ç¿ãå©ããåªè¯ãªãã¥ã¼ããªã¢ã«ã¾ã¨ã
ããããªç½ ããããæããããã«ãæè¿ä¸çªãããªã¨æãã®ãå®è·µå½¢å¼ã®ãã¥ã¼ããªã¢ã«ã§ãã説æãèªã¿ãªãããæãåãããã¨ã§ç解ãæ·±ã¾ãã¾ããç¹ã«æè¿ã®åªè¯ãªãã¥ã¼ããªã¢ã«ã¯ãç°å¢æ§ç¯ä¸è¦ã§ãããå¦ç¿ã§ãããããªå·¥å¤«ããã¦ããã®ã§ãæçã§çµæãåºãã¾ãã
ãç§ã¯ãã®åãé åãããã¦ãã¾ãã¾ããããããããæçã§å¦ç¿é²ããã人ããå¦ç¿ã«è¡ãè©°ã¾ã£ã人ãæ¬ã«é£½ãã¦ãã人ã¯ãã¥ã¼ããªã¢ã«ã試ãã¦ã¿ããã¨ããå§ããã¾ããã人工ç¥è½ã¯äººé¡ãæ» ã¼ããã¨ããæ¬ãèªãã§æ¯ãã¦ãã人ããå®éã«èªåã§æãåããã¦æå 端ã®AIã¢ãã«ãè²ã¦ã¦è¦ãã¨ãããAIçµæ§å¯æãããããã¨ããAIã£ã¦è¶ ã¢ãã ãªãçãä»ã¾ã§ã¨äººå·¥ç¥è½ã«å¯¾ããè¦æ¹ã大ããå¤ããããããã¾ãããã
ããããªäººå·¥ç¥è½ã«å¯¾ããè¦æ¹ãå¤ãããããããªããåªè¯ãã¥ã¼ããªã¢ã«ãç´¹ä»ãã¦ããã¾ããææãªãã®ãå«ã¾ãã¦ãã¾ããã¿ã¤ãã«ã®æå¾ã«ãç¡æã»ææã¨è¨è¼ãã¤ãã¦ããã®ã§åèã«ãã¦ä¸ããã
æ±äº¬å¤§å¦ã®æ¾å°¾ç 究室ãç¡æå ¬éãã¦ãããDeep Learningåºç¤è¬åº§æ¼ç¿ã³ã³ãã³ããï¼ç¡æï¼
ãã人工ç¥è½ã¯äººéãè¶ ããããã§æåãªæ±å¤§ã®æ¾å°¾å çãç¡æã§å ¬éãã¦ããå¦ç¿ã³ã³ãã³ãã§ãã
ãç¡æã§ãããå 容ã¯ã¨ã¦ãè¯ãã®ã§ãããæããããã¯ç°å¢è¨å®ã使ãæ¹ãå ¨ãæ¸ããã¦ããªãç¹ã§ãã以ä¸è¨äºåèã«ãã¦ã¿ã¦ä¸ããã
ãä¸è¨è¨äºã§ãé£ããã¨ããæ¹ã¯ãä»ã®ææã®ãã¥ã¼ããªã¢ã«ã試ãã¦è¦ã¾ããããå°ã£ãããéã§è§£æ±ºããã®ã大人ã§ãï¼
Aidemyï¼ä¸é¨ç¡æï¼
ããã£ã¼ãã©ã¼ãã³ã°ãå¦ç¿ã§ããWebãµã¼ãã¹ãAidemyããã§ããå¦ç¿ç°å¢ã¾ã§æä¾ããã¦ããã®ã§ãæ¬å½ã«10ç§ã§å¦ç¿ãéå§ã§ãã¾ããåºæ¬çã«ææã§ãããå¤ãã®ã»ããã¼ãªã©ã«æ¯è¼ããã¨æ¬å½ã«æ ¼å®ã§ãããä¸é¨ã¯ç¡æã§å¦ç¿ã§ããã®ã§ãç¡æã®ã³ã¼ã¹ã ãã§ã試ãã¦ã¿ã価å¤ãããã¨æãã¾ãã
ãã¹ãå¥ã«è³ãããã£ãããã¹ãããã¦ãããããããªãã§ããããï¼
PFNã®ã¨ã³ã¸ãã¢ãããå ¬éãã¦ãããã¥ã¼ããªã¢ã«ï¼ç¡æï¼
ãä»ãããã¡ã¯æ¥æ¬ã®AIãã³ãã£ã¼ãPFNãã®ã¨ã³ã¸ãã¢ãããå ¬éãã¦ãããã¥ã¼ããªã¢ã«ã§ãã以åä¸è¨ã®è¨äºã§ãã¼ã«ã«PCã§ã®ã»ããã¢ããæ¹æ³ã¨ä½¿ãæ¹å«ãã¦ç´¹ä»ãã¾ããã使ç¨ãã¦ãããã£ã¼ãã©ã¼ãã³ã°ã®ãã¬ã¼ã ã¯ã¼ã¯ã¯PFNéçºã®Chainerã¨ãªãã¾ããChainer使ãã®æ¹ã¯ãã¡ãããä»ã®ãã¬ã¼ã ã¯ã¼ã¯ã«ãå ±éãããã£ã¼ãã©ã¼ãã³ã°ã®å¦ç¿ã«é¢ããä¸å¯§ãªè§£èª¬ãããã®ã§ãChainer使ããªã人ã§ãã¨ã¦ãåèã«ãªãã¨æãã¾ãã
ãææ°ã®ä»¥ä¸ã®ãã¥ã¼ããªã¢ã«ã§ã¯ããGoogle Colabratoryãã¨ããGoogleæä¾ã®ç¡æã¯ã©ã¦ããµã¼ãã¹ä¸ã§ãã¥ã¼ããªã¢ã«ãåããæ¹æ³ãè¨è¼ããã¦ãã¾ãã
ããã ãå ¨ä½çã«ããç¨åº¦åãã£ã¦ãã人åãã®è§£èª¬ãªã®ã§ãåå¦è ã«ã¯é£ããã¨ãããããããããã¾ãããã¨æããããPFNãããChainerã®ãã¥ã¼ããªã¢ã«ã®æ±ºå®çã¨ãè¨ããå 容ã以ä¸ãµã¤ãã§å ¬éãã¦ä¸ãã£ã¦ãã¾ãã
ãPythonã®åºç¤ããå§ã¾ããæ©æ¢°å¦ç¿ã®åºç¤ãæ©æ¢°å¦ç¿é¢é£ã®Pythonã®ä¸»è¦ã©ã¤ãã©ãªï¼Numpy, Matplotlibçï¼ã®ä½¿ãæ¹ããã£ã¼ãã©ã¼ãã³ã°ï¼Chainerï¼ã®åºç¤ããç»åèªèã®å¿ç¨ã¾ã§ãã¾ã ãå¿ç¨ã«é¢ãã¦ã¯ä½æä¸ã®ããã§ãããé常ã«ä»å¾ãæå¾ ã§ããå 容ã§ãã大å¦çã対象ã«ãã¦ããããã§ãããåºç¤ããå¦ã³ãã社ä¼äººã«ãåèã«ãªãã¨æãã¾ãã
Kaggleã®ãã¥ã¼ããªã¢ã«ï¼ææï¼
ãã«ã¬ã¼ã¡ãããããæænoteã§å ¬éãã¦ããæ©æ¢°å¦ç¿ã³ã³ãKaggleã®ãã¥ã¼ããªã¢ã«ã§ããKaggleã¯ããªãã¬ããªäººãå¤ãå°è±¡ãªã®ã§ãç§ã¯ã¨ã¦ã足ãè¸ã¿å ¥ããããªãã®ã§ãããã¬ãå¢ã«ãªãããæ¹ã¯æ¯éï¼
ãKaggleèªä½ã¯ç¡æãªã®ã§ãã¨ããããç¡æã§Kaggleããã人ã¯ä»¥ä¸è¨äºåç §ä¸ãã
ç§éDeep Learningï¼ææï¼
ããã£ã¼ãã©ã¼ãã³ã°ã®è§£èª¬ããã¯ãã¾ããåå¼·æ¹æ³ã¨ã¤ã¾ã¥ããã¤ã³ããããã«ã¯ãGoogle Colaboratoryãã使ç¨ãã¦ãã¥ã¼ããªã¢ã«ã§æãå¦ç¿ããªããå¦ã¶æ¹æ³ã¾ã§å¹ åºãçãè¾¼ã¾ãã¦ããåªè¯ãªé»åæ¸ç±ã§ããå ã»ã©ç´¹ä»ãããæ±å¤§ã®æ¾å°¾ç ã®ãã¥ã¼ããªã¢ã«ãGoogle Colaboratoryã§åããæ¹æ³ãä¸å¯§ã«è§£èª¬ããã¦ãã¾ããä½è ã®æ¹ã¯ãå°é家ã§ã¯ãªãç¬å¦ã§å¦ç¿é²ããã¨ãããã¨ã§ãåå¦è ã«è¿ãè¦ç¹ã§æ¸ããã¦ãããé常ã«åãããããã§ããåãç¬å¦è ã¨ãã¦ãå 容ã«å ±æããã¨ãããå¤ãã£ãã§ãã
ãçè«ããå®è·µã¾ã§ãã©ã³ã¹è¯ããå¤ãã®å¸è²©æ¸ã«æ¯ã¹ã¦å ¨ãéè²ãªãã³ã¹ãããã©ã¼ãã³ã¹ï¼ã¨ãããå¤ãã®ãã®ãåé§ãã¦ããï¼ã¨æãã¾ãã絶対買ãã§ããã
ã以ä¸ã®Qiitaè¨äºãç¡æã§å ¬éããã¦ãã¾ããããããæ´ã«10å以ä¸æ¿å¯ã«ããå 容ã«ãªã£ã¦ãã¾ãããéåºãã¦ããã¹ãå 容ã¨æãã¾ãã
fast.ai ï¼ç¡æï¼
ãç¡æã§ãã£ã¼ãã©ã¼ãã³ã°ãå¦ã¹ãæµ·å¤ã®ææã§ãã使ç¨ãã¦ãããã¬ã¼ã ã¯ã¼ã¯ï¼fastaiï¼ã¯PyTorchãã¼ã¹ã®ããã§ããåç»ã¨Jupyter Notebookå½¢å¼ã®ãã¥ã¼ããªã¢ã«ã§å¦ç¿ã§ããããã§ãã
ãè±èªã§ç¸å½ã®ããªã¥ã¼ã ããããããªããã¼ãã«é«ãã®ã§ãããhiromisããã¨ããæ¹ããè¬ç¾©ãã¾ã¨ããè³æãGitHubã«ã¢ãããã¦ãã ãã£ã¦ãã¾ãããã¡ãæç¨ã§ãï¼
ãã¾ããç§éDeep Learningã®ä½è tomo-makesããããfast.aiã®GitHubã®Jupyter Notebookãforkãã¦æ¥æ¬èªè¨³ãã¦ãã ãã£ã¦ããããã¡ããé常ã«å©ãã«ãªãããã§ãã
ããããã®æ å ±ã¯ãtomo-makesããã®ãã¤ã¼ãã§ç¥ãã¾ãããæè¬ã§ãã
ãã¾ã ãã¡ãã¯åãçµãã¦ãã¾ããããPyTorchã«ãå°ãèå³ããã£ãã®ã§ãæ°ã«ãªãã¨ããã ãã§ã試ãã¦ã¿ãããã¨æã£ã¦ãã¾ãã
https://t.co/rumudQ4iTEã®å®è·µDLè¬åº§ãhiromisæ°ã®ãã¼ãããããã¾ã¨ã¾ã£ã¦ãã¾ãã(æ¨å¹´ãæ¸ããã¦ããæ¹?) 2æéx7åã®æéã¯ç¢ºä¿ã§ããªããè¬ç¾©åç»ã§è±èªã¯è¿½ãã¥ãããè¬åº§å¾ã®å¾©ç¿ã«ãããããåç»ã¨åçã®æ å ±éããããããhttps://t.co/lDvPF6cEMO
— tomo-makesï¼ ç§éDeep Learningæ¬ã¯BOOTH㧠(@tomo_makes) 2019å¹´1æ27æ¥
è¬åº§æ¬ä½ã¯ãhttps://t.co/PhxwMopPbt
fastaiã®2019å¹´çæ°è¬åº§ã試ããã¦ããåJupyter notebookã®ãã£ããæ¥æ¬èªè¨³çãä½ãã¾ãããhttps://t.co/D1moIcRVbJ
— tomo-makesï¼ ç§éDeep Learningæ¬ã¯BOOTH㧠(@tomo_makes) 2019å¹´1æ26æ¥
Colaboratoryä¸ã§Python3/GPUã«ã¼ãã«ãé¸ã³ã `!curl https://t.co/KwErYoGZsS | bash` ãå®è¡ããã¨ãå ¨ã¦åããã¾ããColabã¯PyTorchãããã©ã«ãæè¼ã«ãªã£ããããã https://t.co/0L7JXexuam
ã¾ã¨ã
ããã£ã¼ãã©ã¼ãã³ã°ã®åªè¯ãªãªã¹ã¹ã¡ãã¥ã¼ããªã¢ã«ãç´¹ä»ãã¾ãããæãåããã®ãä¸çªã£ã¦ãã£ã¼ãã©ã¼ãã³ã°ãããããè¨ã£ã¦ããã®ã§ããã£ãã身ã«ä»ããã人ã¯ããã¥ã¼ããªã¢ã«ãè¿éãããªãããªã¨æãã¾ããæ¬è¨äºãåèã«æ¯éè²ã ãããã¦ã¿ã¦ãã ããã
æ¸ç±ã®å®£ä¼ãããããå çã®ã¨ã«ãã楽ããAIèªä½æ室ã
ãAIã®åå¿è åãæ¬ãããããå çã®ã¨ã«ãã楽ããAIèªä½æ室ããå·çãã¾ãããGoogle Colaboratoryã使ã£ã¦ãå®éã«åãããªããAIãå¦ã¹ãæ¬ã«ãªã£ã¦ãã¾ãã
ãæç³»ã§AIç´ äººã®ç§ã®å¦»ããå®è·µã§ãã¦ç解ãæ·±ãããããããªãåå¦è åãã®å 容ã«ãªã£ã¦ãã¾ãã以ä¸ã«ç´¹ä»è¨äºãæ¸ãã¦ãã¾ãã®ã§ãèå³ããæ¹ã¯æ¯éããããã ããã
ãã®ä»ã®åªè¯ãªãªã³ã¯
ãç´¹ä»ããããªãã£ãåªè¯ãªæ å ±ã¸ã®ãªã³ã¯ã§ããå¥ã®æ©ä¼ã«ç´¹ä»ãããã¨æãã¾ãã
GitHub - ageron/handson-ml2: A series of Jupyter notebooks that walk you through the fundamentals of Machine Learning and Deep Learning in Python using Scikit-Learn, Keras and TensorFlow 2.
https://github.com/yoyoyo-yo/DeepLearningMugenKnock
GitHub - nagaoka-ai-innovationhub/basics-of-image-recognition-with-cnn
GitHub - tugstugi/dl-colab-notebooks: Try out deep learning models online on Google Colab
The latest in Machine Learning | Papers With Code
Dive into Deep Learning — Dive into Deep Learning 1.0.0-beta0 documentation
GitHub - kuzand/Computer-Vision-Video-Lectures: A curated list of free, high-quality, university-level courses with video lectures related to the field of Computer Vision.
[2009.05673] Applications of Deep Neural Networks with Keras
【Deep Learning研修(発展)】少データ・少ラベル学習 - YouTube
Releases · d2l-ai/d2l-en · GitHub
GitHub - suhara/cis6930-fall2021: Course materials for Fall 2021 "CIS6930 Topics in Computing for Data Science" at New College of Florida
【2023年版】機械学習の日本語無料学習教材まとめ - Qiita
é¢é£è¨äº
å¤æ´å±¥æ´
- 2020/12/20 AIæ¸ç±ã®å·çã«é¢ãã¦è¿½è¨
- 2020/06/19 é¢é£è¨äºè¿½è¨
- 2020/02/24 ãªã³ã¯è¿½è¨
- 2020/01/24 åªè¯ãªãªã³ã¯ã追è¨
- 2019/11/20 é¢é£è¨äºã»åèãªã³ã¯è¿½è¨
- 2019/04/10 PFNããã®Chainerãã¥ã¼ããªã¢ã«ã«é¢ãã¦è¿½è¨
- 2019/01/28 fast.aiã«é¢ãã¦è¿½è¨