Introduction of Deep Reinforcement Learning, which was presented at domestic NLP conference. è¨èªå¦çå¦ä¼ç¬¬24å年次大ä¼(NLP2018) ã§ã®è¬æ¼è³æã§ãã http://www.anlp.jp/nlp2018/#tutorialRead less
å¼·åå¦ç¿ã®ä¸ææ³ã§ããQ-learning ã¨ãã£ã¼ããã¥ã¼ã©ã«ããããçµã¿åããã Deep Q Networkãé称DQNã使ã£ã¦åç«æ¯åã®æ¯ãä¸ãåé¡ã解決ãã¦ã¿ã¾ãã åé¡è¨å® ãåç«æ¯åã®æ¯ãä¸ãåé¡ãã¨ããã®ã¯ãä»åã¯ããããåé¡è¨å®ã§ãã ã¾ã空ä¸ã«éæ¢ããã¢ã¼ã¿ããã£ã¦ãã¢ã¼ã¿è»¸ã«æ£ã®ä¸ç«¯ãã¤ãªãã£ã¦ãã¾ããæ£ã¯ä¸å¿ã«è³ªéãéä¸ãã¦ãã¦åæ§$\infty$ã§å¤ªã0ã®ãããããæ£ã§ããåæç¶æ ã§ã¯æ£ã¯éåã«ãããã£ã¦ä¸åãã«ã¶ãä¸ãã£ã¦ãã¾ãããã®ç¶æ ããæ¯ãåãæ¯ãä¸ãã¦åç«ç¶æ ã§éæ¢ããã¦ãã ãããã¨ããåé¡ã§ããå¤ãããå¶å¾¡å·¥å¦ã§ã¯ãæ¯ãä¸ãç¨ã¨éæ¢ç¨ã«å¥è¨è¨ãããã³ã³ããã¼ã©ã2ã¤ç¨æãã¦åãæ¿ãããªã©ãéç·å½¢è¦ç´ ãå«ãã³ã³ããã¼ã©ãç¨ãã¦å¯¾å¦ãããã¨ã«ãªãã¾ããããããã£ããã¨ãªãã§ããã©ããããããã§ãã ä»åã¯ãã¢ã¼ã¿ã¯å³ãå·¦ã«ä¸å®ãã«ã¯ã®å転ããã§ããªããã¨ã
ã¡ã³ããã³ã¹
ãç¥ãã
é害
ãªãªã¼ã¹ãé害æ å ±ãªã©ã®ãµã¼ãã¹ã®ãç¥ãã
ææ°ã®äººæ°ã¨ã³ããªã¼ã®é ä¿¡
å¦çãå®è¡ä¸ã§ã
j次ã®ããã¯ãã¼ã¯
kåã®ããã¯ãã¼ã¯
lãã¨ã§èªã
eã³ã¡ã³ãä¸è¦§ãéã
oãã¼ã¸ãéã
{{#tags}}- {{label}}
{{/tags}}