{"html":"<iframe src=\"https://hatenablog-parts.com/embed?url=https%3A%2F%2Ftadaoyamaoka.hatenablog.com%2Fentry%2F2020%2F04%2F13%2F093651\" title=\"Agent57: Outperforming the Atari Human Benchmark\u3092\u8aad\u3080 \u305d\u306e\uff14 - TadaoYamaoka\u306e\u958b\u767a\u65e5\u8a18\" class=\"embed-card embed-blogcard\" scrolling=\"no\" frameborder=\"0\" style=\"display: block; width: 100%; height: 190px; max-width: 500px; margin: 10px 0px;\"></iframe>","provider_name":"Hatena Blog","blog_title":"TadaoYamaoka\u306e\u958b\u767a\u65e5\u8a18","description":"\u5b9f\u9a13 \u3053\u306e\u7ae0\u3067\u306f\u3001\u307e\u305a\u5b9f\u9a13\u306e\u30bb\u30c3\u30c8\u30a2\u30c3\u30d7\u306b\u3064\u3044\u3066\u8aac\u660e\u3059\u308b\u3002 NGU\u306b\u7d9a\u3044\u3066\u3001Agent57\u306f\u30b5\u30a4\u30baN = 32\u306e\u4fc2\u6570\u306e\u30d5\u30a1\u30df\u30ea\u30fc\u3092\u4f7f\u7528\u3059\u308b\u3002 \u5272\u5f15\u306e\u9078\u629e\u306fNGU\u306e\u305d\u308c\u3068\u306f\u7570\u306a\u308a\u30010.99\u304b\u30890.9999\u306e\u7bc4\u56f2\u306e\u3088\u308a\u9ad8\u3044\u5024\u3092\u53ef\u80fd\u306b\u3059\u308b\uff08\u8a73\u7d30\u306b\u3064\u3044\u3066\u306f\u4ed8\u9332G.1\u3092\u53c2\u7167\uff09\u3002 \u30e1\u30bf\u30b3\u30f3\u30c8\u30ed\u30fc\u30e9\u30fc\u306f\u3001\u30a2\u30af\u30bf\u30fc\u306b\u03c4= 160\u30a8\u30d4\u30bd\u30fc\u30c9\u3068\u03b5= 0.5\u306e\u30a6\u30a3\u30f3\u30c9\u30a6\u30b5\u30a4\u30ba\u3092\u4f7f\u7528\u3057\u3001\u03c4= 3600\u30a8\u30d4\u30bd\u30fc\u30c9\u3068\u03b5= 0.01\u306e\u30a6\u30a3\u30f3\u30c9\u30a6\u30b5\u30a4\u30ba\u3092\u4f7f\u7528\u3059\u308b\u3002 \u305d\u306e\u4ed6\u306e\u30cf\u30a4\u30d1\u30fc\u30d1\u30e9\u30e1\u30fc\u30bf\u306f\u3059\u3079\u3066\u3001Atari\u30d5\u30ec\u30fc\u30e0\u306e\u6a19\u6e96\u7684\u306a\u524d\u51e6\u7406\u3092\u542b\u3081\u3001NGU\u306e\u30cf\u30a4\u30d1\u30fc\u30d1\u30e9\u30e1\u30fc\u30bf\u3068\u540c\u3058\u3067\u3042\u308b\u3002 \u4f7f\u7528\u3059\u308b\u30cf\u30a4\u30d1\u30fc\u30d1\u30e9\u30e1\u30fc\u30bf\u3068\u524d\u51e6\u7406\u306e\u8a73\u7d30\u306b\u2026","width":"100%","published":"2020-04-13 09:36:51","author_name":"TadaoYamaoka","height":"190","provider_url":"https://hatena.blog","blog_url":"https://tadaoyamaoka.hatenablog.com/","categories":["\u5f37\u5316\u5b66\u7fd2","Agent57"],"version":"1.0","url":"https://tadaoyamaoka.hatenablog.com/entry/2020/04/13/093651","title":"Agent57: Outperforming the Atari Human Benchmark\u3092\u8aad\u3080 \u305d\u306e\uff14","image_url":"https://chart.apis.google.com/chart?cht=tx&chl=%5C%7B%28%5Cbeta_j%2C%5Cgamma_j%29%5C%7D_%7Bj%3D0%7D%5E%7BN-1%7D","type":"rich","author_url":"https://blog.hatena.ne.jp/TadaoYamaoka/"}