\n1. \uac15\ud654\ud559\uc2b5\uc758 \uc774\ud574<\/h2>\n
\n \uac15\ud654\ud559\uc2b5(Reinforcement Learning, RL)\uc740 \uc778\uacf5\uc9c0\ub2a5\uc758 \ud55c \ubd84\uc57c\ub85c, \uc5d0\uc774\uc804\ud2b8\uac00 \ud658\uacbd\uacfc \uc0c1\ud638\uc791\uc6a9 \ud558\uba70 \ucd5c\uc801\uc758 \ud589\ub3d9\uc744 \ud559\uc2b5\ud558\ub294 \uc54c\uace0\ub9ac\uc998\uc785\ub2c8\ub2e4.
\n \uc774 \uacfc\uc815\uc5d0\uc11c \uc5d0\uc774\uc804\ud2b8\ub294 \ubcf4\uc0c1\uc744 \uc5bb\uae30 \uc704\ud574 \uc5ec\ub7ec \uac00\uc9c0 \ud589\ub3d9\uc744 \uc2dc\ub3c4\ud558\uace0 \uacb0\uacfc\ub97c \ubc14\ud0d5\uc73c\ub85c \ud559\uc2b5\ud569\ub2c8\ub2e4.
\n \uac15\ud654\ud559\uc2b5\uc758 \ub300\ud45c\uc801\uc778 \ud2b9\uc9d5\uc740 \uc2dc\ud5d8\uacfc \uc624\ub958(trial and error)<\/strong> \ud559\uc2b5 \ubc29\uc2dd\uacfc \uc7a5\uae30\uc801\uc778 \ubcf4\uc0c1(long-term reward)<\/strong> \ucd5c\uc801\ud654\uc785\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\n2. \uac15\ud654\ud559\uc2b5\uc758 \uc8fc\uc694 \uac1c\ub150<\/h2>\n
\n \uac15\ud654\ud559\uc2b5\uc744 \uc774\ud574\ud558\uae30 \uc704\ud574\uc11c \uc54c\uc544\uc57c \ud560 \uc8fc\uc694 \uac1c\ub150\uc740 \ub2e4\uc74c\uacfc \uac19\uc2b5\ub2c8\ub2e4:\n <\/p>\n
\n- \uc5d0\uc774\uc804\ud2b8(Agent)<\/strong>: \ud658\uacbd\uacfc \uc0c1\ud638\uc791\uc6a9\ud558\uba70 \ud559\uc2b5\ud558\ub294 \uc8fc\uccb4\uc785\ub2c8\ub2e4.<\/li>\n
- \ud658\uacbd(Environment)<\/strong>: \uc5d0\uc774\uc804\ud2b8\uac00 \uc0c1\ud638\uc791\uc6a9\ud558\ub294 \uc138\uc0c1\uc785\ub2c8\ub2e4. \uc0c1\ud0dc\uc640 \ubcf4\uc0c1\uc744 \uc81c\uacf5\ud569\ub2c8\ub2e4.<\/li>\n
- \uc0c1\ud0dc(State)<\/strong>: \uc5d0\uc774\uc804\ud2b8\uac00 \ud658\uacbd\uc5d0\uc11c \ubd88\ub7ec\uc628 \uc815\ubcf4\uc785\ub2c8\ub2e4. \ud604\uc7ac \uc0c1\ud669\uc744 \ub098\ud0c0\ub0c5\ub2c8\ub2e4.<\/li>\n
- \ud589\ub3d9(Action)<\/strong>: \uc5d0\uc774\uc804\ud2b8\uac00 \uc8fc\uc5b4\uc9c4 \uc0c1\ud0dc\uc5d0\uc11c \uc120\ud0dd\ud560 \uc218 \uc788\ub294 \ubaa8\ub4e0 \uac00\ub2a5\uc131\uc785\ub2c8\ub2e4.<\/li>\n
- \ubcf4\uc0c1(Reward)<\/strong>: \uc5d0\uc774\uc804\ud2b8\uc758 \ud589\ub3d9\uc5d0 \ub300\ud55c \ud53c\ub4dc\ubc31\uc73c\ub85c, \ud589\ub3d9\uc758 \uacb0\uacfc\uc5d0 \ub530\ub77c \uc8fc\uc5b4\uc9c0\ub294 \uac12\uc785\ub2c8\ub2e4.<\/li>\n
- \uc815\ucc45(Policy)<\/strong>: \ud2b9\uc815 \uc0c1\ud0dc\uc5d0\uc11c \uc5b4\ub5a4 \ud589\ub3d9\uc744 \ucde8\ud560\uc9c0\ub97c \uacb0\uc815\ud558\ub294 \ud568\uc218\uc785\ub2c8\ub2e4.<\/li>\n
- \uac00\uce58\ud568\uc218(Value Function)<\/strong>: \ud2b9\uc815 \uc0c1\ud0dc\uc5d0\uc11c \uae30\ub300\ub418\ub294 \ubcf4\uc0c1\uc758 \ud569\uc785\ub2c8\ub2e4.<\/li>\n<\/ul>\n<\/section>\n\n
3. \ucd94\ucc9c \ub3c4\uc11c<\/h2>\n
\uac15\ud654\ud559\uc2b5\uc5d0 \uad00\ud55c \uc774\ub860\uacfc \uc2e4\uc81c\ub97c \uae4a\uc774 \uc788\uac8c \ubc30\uc6b0\uae30 \uc704\ud574 \ub2e4\uc74c\uacfc \uac19\uc740 \ucd94\ucc9c \ub3c4\uc11c\ub97c \uace0\ub824\ud574\ubcf4\uc138\uc694:<\/p>\n
3.1. “Reinforcement Learning: An Introduction” – Richard S. Sutton, Andrew G. Barto<\/h3>\n
\n \uc774 \ucc45\uc740 \uac15\ud654\ud559\uc2b5\uc758 \uae30\ucd08\ub97c \ub2e4\uc9c8 \uc218 \uc788\ub294 \uac00\uc7a5 \uad8c\uc704 \uc788\ub294 \uc790\ub8cc \uc911 \ud558\ub098\uc785\ub2c8\ub2e4.
\n \uae30\ubcf8 \uac1c\ub150\ubd80\ud130 \uace0\uae09 \uae30\ubc95\uae4c\uc9c0 \uccb4\uacc4\uc801\uc73c\ub85c \uc124\uba85\ud558\uace0 \uc788\uc5b4, \uc804\uacf5\uc790\ubfd0\ub9cc \uc544\ub2c8\ub77c \ube44\uc804\uacf5\uc790\ub3c4 \uc27d\uac8c \uc774\ud574\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n
3.2. “Deep Reinforcement Learning Hands-On” – Maxim Lapan<\/h3>\n
\n \uc774 \ucc45\uc740 \ub525\ub7ec\ub2dd\uacfc \uac15\ud654\ud559\uc2b5\uc744 \uacb0\ud569\ud55c \uc2e4\uc81c \uc0ac\ub840\ub97c \ub2e4\ub8f9\ub2c8\ub2e4.
\n \ud30c\uc774\uc36c\uacfc PyTorch\ub97c \uc774\uc6a9\ud55c \ub2e4\uc591\ud55c \uc608\uc81c\ub97c \ud1b5\ud574 \uc2e4\ub825\uc744 \uc313\uc744 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n
3.3. “Algorithms for Reinforcement Learning” – Csaba Szepesv\u00e1ri<\/h3>\n
\n \uc774 \ucc45\uc740 \uac15\ud654\ud559\uc2b5\uc758 \uc218\ud559\uc801 \uae30\ucd08\uc640 \uc54c\uace0\ub9ac\uc998\uc744 \uc2ec\ub3c4 \uc788\uac8c \ub2e4\ub8f9\ub2c8\ub2e4.
\n \ubcf4\ucda9 \uc790\ub8cc\ub85c \ud65c\uc6a9\ud558\uba74 \uc774\ub860\uc744 \uc880 \ub354 \uacac\uace0\ud788 \ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\n4. \uc628\ub77c\uc778 \uac15\uc758 \ubc0f \ud29c\ud1a0\ub9ac\uc5bc<\/h2>\n
\n \uc628\ub77c\uc778 \uc790\ub8cc\ub4e4\uc740 \uc2dc\uac04\uacfc \uc7a5\uc18c\uc5d0 \uad6c\uc560\ubc1b\uc9c0 \uc54a\uace0 \ud559\uc2b5\ud560 \uc218 \uc788\ub294 \uc7a5\uc810\uc744 \uc81c\uacf5\ud569\ub2c8\ub2e4. \ub2e4\uc591\ud55c \uac15\uc758 \ud50c\ub7ab\ud3fc\uc5d0\uc11c \uc81c\uacf5\ud558\ub294 \uac15\ud654\ud559\uc2b5 \uad00\ub828 \uac15\uc758\ub97c \ucd94\ucc9c\ud569\ub2c8\ub2e4:\n <\/p>\n
4.1. Coursera<\/h3>\n
\n Coursera\ub294 \uc2a4\ud0e0\ud3ec\ub4dc \ub300\ud559\uad50\uc758 “Reinforcement Learning Specialization” \ubc0f \ub2e4\ub978 \uc720\uba85 \ub300\ud559\uc758 \uac15\uc758\ub97c \uc81c\uacf5\ud569\ub2c8\ub2e4.
\n \uc774 \uac15\uc758\ub294 \uae30\ucd08\ubd80\ud130 \uace0\uae09 \uacfc\uc815\uc744 \ud3ec\uad04\uc801\uc73c\ub85c \ub2e4\ub8f9\ub2c8\ub2e4.\n <\/p>\n
4.2. edX<\/h3>\n
\n edX \ud50c\ub7ab\ud3fc\uc5d0\uc11c\ub3c4 MIT, Harvard \ub4f1\uc5d0\uc11c \uc81c\uacf5\ud558\ub294 \uac15\ud654\ud559\uc2b5 \uad00\ub828 \uac15\uc758\ub97c \ub4e4\uc744 \uc218 \uc788\uc2b5\ub2c8\ub2e4.
\n \uc774 \uac15\uc758\ub4e4\uc740 \uc8fc\ub85c \uc5f0\uad6c \uc911\uc2ec\uc758 \ub0b4\uc6a9\uc744 \ub2e4\ub8f9\ub2c8\ub2e4.\n <\/p>\n
4.3. Udacity<\/h3>\n
\n Udacity\ub294 “Deep Reinforcement Learning Nanodegree” \ud504\ub85c\uadf8\ub7a8\uc744 \uc81c\uacf5\ud558\uc5ec \uc2e4\ubb34 \uc911\uc2ec\uc758 \uac15\uc758\ub97c \uc81c\uacf5\ud569\ub2c8\ub2e4.
\n \uc2e4\uc2b5 \ud504\ub85c\uc81d\ud2b8\ub3c4 \ud3ec\ud568\ub418\uc5b4 \uc788\uc5b4 \uc2e4\uc9c8\uc801\uc778 \uacbd\ud5d8\uc744 \uc313\uc744 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\n5. \uac15\ud654\ud559\uc2b5 \ucee4\ubba4\ub2c8\ud2f0<\/h2>\n
\n \uac15\ud654\ud559\uc2b5\uc5d0 \ub300\ud55c \ud1a0\ub860\uacfc \uc815\ubcf4\ub97c \ub098\ub204\uae30 \uc704\ud55c \ucee4\ubba4\ub2c8\ud2f0\ub3c4 \uc911\uc694\ud569\ub2c8\ub2e4. \ub2e4\uc74c\uc758 \ud50c\ub7ab\ud3fc\uc5d0\uc11c \ub2e4\uc591\ud55c \ud1a0\ub860\uacfc \ud559\uc2b5 \uc790\ub8cc\ub97c \ucc3e\uc744 \uc218 \uc788\uc2b5\ub2c8\ub2e4:\n <\/p>\n
5.1. GitHub<\/h3>\n
\n GitHub\uc5d0\ub294 \uac15\ud654\ud559\uc2b5 \uad00\ub828 \ud504\ub85c\uc81d\ud2b8\uc640 \ucf54\ub4dc \uc0d8\ud50c\uc774 \ub2e4\uc218 \uc874\uc7ac\ud569\ub2c8\ub2e4.
\n \uacf5\uac1c\ub41c \ub808\ud3ec\uc9c0\ud1a0\ub9ac\uc5d0\uc11c \ub2e4\uc591\ud55c \uc54c\uace0\ub9ac\uc998\uacfc \uc2e4\uc81c \uad6c\ud604 \ubc29\ubc95\uc744 \ubc30\uc6b8 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n
5.2. Reddit<\/h3>\n
\n Reddit\uc758 r\/reinforcementlearning \ucee4\ubba4\ub2c8\ud2f0\ub294 \uac15\ud654\ud559\uc2b5 \uad00\ub828 \ucd5c\uc2e0 \uc5f0\uad6c \ubc0f \uc801\uc6a9 \uc0ac\ub840\ub97c \uacf5\uc720\ud558\ub294 \uc7a5\uc18c\uc785\ub2c8\ub2e4.
\n \uc5ec\ub7ec\ubd84\uc740 \uc9c8\ubb38\uc744 \ud558\uace0, \ub2e4\ub978 \ud559\uc2b5\uc790\ub4e4\uacfc \uacbd\ud5d8\uc744 \ub098\ub20c \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n
5.3. Stack Overflow<\/h3>\n
\n \uac1c\ubc1c \uc911 \ubb38\uc81c\uac00 \ubc1c\uc0dd\ud588\uc744 \ub54c, Stack Overflow\uc5d0\uc11c \uc9c8\ubb38\uc744 \ud558\uace0 \ub2f5\ubcc0\uc744 \ucc3e\uc544 \ubb38\uc81c\ub97c \ud574\uacb0\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.
\n \uc5ec\ub7ec \uacbd\ud5d8 \ub9ce\uc740 \uac1c\ubc1c\uc790\ub4e4\uc5d0\uac8c \ub3c4\uc6c0\uc744 \ubc1b\uc744 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\n6. \ud559\uc2b5\uc744 \uc704\ud55c \ud301<\/h2>\n
\n \uac15\ud654\ud559\uc2b5\uc740 \ubcf5\uc7a1\ud55c \ubd84\uc57c\uc9c0\ub9cc, \uc544\ub798\uc640 \uac19\uc740 \ubc29\ubc95\uc73c\ub85c \ud559\uc2b5\uc744 \uc9c4\ud589\ud558\uba74 \uc88b\uc2b5\ub2c8\ub2e4:\n <\/p>\n
\n- \uae30\ucd08 \uc774\ub860\uc744 \uc219\uc9c0\ud558\ub77c<\/strong>: \uac15\ud654\ud559\uc2b5\uc758 \uae30\ucd08 \uac1c\ub150\uacfc \uc54c\uace0\ub9ac\uc998(\uc608: Q-learning, SARSA)\uc744 \uc774\ud574\ud558\ub294 \uac83\uc744 \uc6b0\uc120\uc801\uc73c\ub85c \ud558\uc2ed\uc2dc\uc624.<\/li>\n
- \uc2e4\uc2b5\uc744 \ubcd1\ud589\ud558\ub77c<\/strong>: \ud559\uc2b5\ud55c \uc774\ub860\uc744 \uc2e4\uc81c \ucf54\ub4dc\ub85c \uad6c\ud604\ud558\ub294 \uac83\uc774 \uc911\uc694\ud569\ub2c8\ub2e4. \ub2e4\uc591\ud55c \ud29c\ud1a0\ub9ac\uc5bc\uc744 \ud1b5\ud574 \uc791\uc740 \ud504\ub85c\uc81d\ud2b8\ub97c \uc9c4\ud589\ud574 \ubcf4\uc138\uc694.<\/li>\n
- \ucee4\ubba4\ub2c8\ud2f0\uc5d0 \ucc38\uc5ec\ud558\ub77c<\/strong>: \ub2e4\ub978 \ud559\uc2b5\uc790\ub4e4\uacfc \uacbd\ud5d8\uc744 \ub098\ub204\uace0 \ubb38\uc81c\ub97c \ud574\uacb0\ud558\uae30 \uc704\ud574 \uc801\uadf9\uc801\uc73c\ub85c \ucee4\ubba4\ub2c8\ud2f0\uc5d0 \ucc38\uc5ec\ud558\uc138\uc694.<\/li>\n
- \uc5f0\uad6c \ub17c\ubb38\uc5d0 \ub3c4\uc804\ud558\ub77c<\/strong>: \ucd5c\uc2e0 \uc5f0\uad6c\ub97c \uc0b4\ud3b4\ubcf4\uba70 \uc0c8\ub85c\uc6b4 \uc54c\uace0\ub9ac\uc998\uacfc \uc811\uadfc \ubc29\ubc95\uc5d0 \ub300\ud574 \ubc30\uc6cc\ubcf4\uc138\uc694. arXiv\ub97c \uc774\uc6a9\ud558\uba74 \ucd5c\uc2e0 \ub17c\ubb38\uc744 \uc27d\uac8c \ucc3e\uc744 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/li>\n<\/ul>\n<\/section>\n