https://www.tensorflow.org/tutorials/reinforcement_learning/actor_critic


Playing CartPole with the Actor-Critic method ย |ย  TensorFlow CorePlaying CartPole with the Actor-Critic method ย |ย  TensorFlow Corewww.tensorflow.org




https://wikidocs.net/172977




์œ„ ๋งํฌ์—์„œ ๊ฐ•ํ™”ํ•™์Šต ์˜ˆ์ œ ์ฝ”๋“œ ๊ณต๋ถ€์ค‘์ž„


๋ฐธ๋ฅ˜ ๋„คํŠธ์›Œํฌ๋ž‘ ์ •์ฑ… ๋„คํŠธ์›Œํฌ ์†์‹ค๊ฐ’์ด๋ž‘ย ๊ฒฝ์‚ฌ ๊ณ„์‚ฐย ๊ฐ™์€ ํŒŒ๋ผ๋ฏธํ„ฐ ์กฐ์ ˆ์„ย ์ง์ ‘ ์ฝ”๋“œํ•œ์ค„์”ฉ ๋‹ค ์ž‘์„ฑํ•ด์„œ ๊ตฌํ˜„ํ–ˆ๋˜๋ฐ


๊ทธ๋ƒฅ fit ํ•จ์ˆ˜ ๋งŒ ๋Œ๋ฆฌ๋ฉด ์•ˆ๋จ?


์†์‹ค ํ•จ์ˆ˜๋„ ๊ทธ๋ƒฅ mse๋ž‘ ํฌ๋กœ์Šค ์—”ํŠธ๋กœํ”ผ ์“ฐ๊ณ ์žˆ๊ณ 


gradient ๋ถ€๋ถ„๋„ ํŠน๋ณ„ํ•œ์ ์„ ๋ชจ๋ฅด๊ฒ ๋„ค


์ €๋ ‡๊ฒŒํ•ด์•ผย ํ–‰๋™ํ•œ๋ฒˆ ํ•  ๋•Œ๋งˆ๋‹ค ํŒŒ๋ผ๋ฏธํ„ฐ ์ˆ˜์ •์ด ๊ฐ€๋Šฅํ•œ๊ฑด๊ฐ€?