{"id":447,"date":"2024-01-11T20:18:16","date_gmt":"2024-01-11T11:18:16","guid":{"rendered":"https:\/\/jaeinlab.com\/?p=447"},"modified":"2024-01-11T20:21:11","modified_gmt":"2024-01-11T11:21:11","slug":"%eb%84%b7%ed%94%8c%eb%a6%ad%ec%8a%a4-%ec%b6%94%ec%b2%9c%ec%8b%9c%ec%8a%a4%ed%85%9c%ec%9d%98-%eb%94%a5%eb%9f%ac%eb%8b%9d-%ec%95%8c%ea%b3%a0%eb%a6%ac%ec%a6%98-%eb%b0%9c%ec%a0%84%ec%82%ac-%ed%8e%98","status":"publish","type":"post","link":"https:\/\/jaeinlab.com\/?p=447","title":{"rendered":"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0)"},"content":{"rendered":"\n
\n

Netflix\uc758 research scientist \uc800\uc790\ub4e4\uc774 \uc791\uc131\ud55c \ub17c\ubb38 “Deep learning for recommender systems: A Netflix case study” \ub9ac\ubdf0\uc785\ub2c8\ub2e4. <\/p>\n\n\n\n

\ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998 \uc5ed\uc0ac\uc5d0 \ud070 \ud55c \ud68d\uc744 \uadf8\uc740 \uae30\uc5c5\uc778 Netflix\uc758 \uc5f0\uad6c\ud300\uc5d0\uc11c \ud37c\ube14\ub9ac\uc2dc\ud55c \ub17c\ubb38\uc778 \ub9cc\ud07c, \uadf8\ub4e4\uc758 struggle\uacfc achievement\ub97c \uac04\uc811\uc801\uc73c\ub85c \ub290\ub07c\uba70 \uac00\ubccd\uac8c(?) \uc77d\uc5b4\ubcf4\uae30\uc5d0 \uc88b\uc740 \uc544\ud2f0\ud074\uc774\uc5c8\uc2b5\ub2c8\ub2e4.<\/p>\n<\/blockquote>\n\n\n

\n
\"\"<\/figure>\n<\/div>\n\n\n

Article Link<\/a><\/p>\n<\/div>\n\n\n\n

Introduction<\/h2>\n\n\n\n

\uc9c0\ub09c \uc218 \ub144\uac04 \ub525\ub7ec\ub2dd\uc740 NLP, CV\uc640 \uac19\uc740 \uc5ec\ub7ec \ubd84\uc57c\uc5d0\uc11c \ube5b\uc744 \ubc1c\ud588\ub294\ub370\uc694, \ucd94\ucc9c\uc2dc\uc2a4\ud15c \ubd84\uc57c\uc5d0\uc11c\ub294 \uc0c1\ub300\uc801\uc73c\ub85c \uadf8 \uc2dc\uae30\uac00 \ub2a6\uac8c \ucc3e\uc544\uc654\ub358 \uac83 \uac19\uc2b5\ub2c8\ub2e4. \uc5b4\ub5a4 \ud0dc\uc2a4\ud06c\uc5d0\uc11c\ub294 \ub525\ub7ec\ub2dd\uc744 \uc801\uc6a9\ud558\ub294 \uac83\uc774 \uc88b\uc740 performance\ub97c \ub0b8 \ubc18\uba74, \uc5b4\ub5a4 \ud0dc\uc2a4\ud06c\uc5d0\uc11c\ub294 \uc5ec\uc804\ud788 \uc624\ud788\ub824 \uc804\ud1b5\uc801\uc778 \ubc29\ubc95\uc774 \ub354 \ud6a8\uacfc\uc801\uc77c \uc815\ub3c4\ub85c, \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc740 NLP\ub098 CV \ubd84\uc57c\uc5d0 \ube44\ud574 \ub09c\ud574\ud55c \ubd80\ubd84\ub4e4\uc774 \ud655\uc2e4\ud788 \uc788\ub294 \uac83 \uac19\uc2b5\ub2c8\ub2e4. \uc774 \ub17c\ubb38\uc5d0\uc11c\ub294 \uc2e4\uc81c Netflix\uc758 \uc5f0\uad6c\uc790\ub4e4\uc774 \ud604\uc5c5\uc5d0\uc11c \ub9c8\uc8fc\uce5c \ub525\ub7ec\ub2dd \uae30\ubc18 \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uac1c\ubc1c\uc758 \uc5b4\ub824\uc6c0\uacfc \uc131\uacfc\uc5d0 \ub300\ud574 \uc9c1\uc811 \uc9c4\ud589\ud55c case study\ub97c \uc18c\uac1c\ud558\uc600\uc2b5\ub2c8\ub2e4. \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc5d0\ub3c4 \ub2e4\uc591\ud55c \ud0dc\uc2a4\ud06c\ub4e4\uc774 \uc874\uc7ac\ud558\ub294\ub370, Netflix\uc5d0\uc11c\ub3c4 \ub9c8\ucc2c\uac00\uc9c0\ub85c \ub2e4\uc591\ud55c \ucd94\ucc9c \ud0dc\uc2a4\ud06c\ub97c \uc11c\ub85c \ub2e4\ub978 \ubaa8\ub378 \uc544\ud0a4\ud14d\ucc98\ub85c \ud480\uc5b4\uc654\ub2e4\uace0 \ud569\ub2c8\ub2e4. \uc774\ubc88 \ub17c\ubb38 \ub9ac\ubdf0\ub97c \ud1b5\ud574, Netflix\uc758 \uc5f0\uad6c\uc790\ub4e4\uc740 \uc5b4\ub5a4 \uc5b4\ub824\uc6c0\ub4e4\uc744 \uadf9\ubcf5\ud558\uc5ec \ub525\ub7ec\ub2dd\uc744 \uc131\uacf5\uc801\uc73c\ub85c \ucd94\ucc9c \ubd84\uc57c\uc5d0 \uc801\uc6a9\ud574\ub0bc \uc218 \uc788\uc5c8\ub294\uc9c0 \uc54c\uc544\ubcf4\ub3c4\ub85d \ud558\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n\n\n\n

Recommendations at Netflix<\/h2>\n\n\n\n

Netflix \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 main goal\uc740 \ub2f9\uc5f0\ud788\ub3c4 \uc0ac\uc6a9\uc790\ub4e4\uc774 \ud074\ub9ad\ud560 \ubc95\ud55c \ucee8\ud150\uce20\ub97c \ucc3e\uc544\uc8fc\ub294 \uac83\uc774\uaca0\uc8e0. \uc0ac\ub78c\ub4e4\uc774 Netflix\ub77c\ub294 \uc11c\ube44\uc2a4\ub97c \uc120\ud0dd\ud558\uace0, \uba64\ubc84\uc27d\uc744 \uc720\uc9c0\ud558\ub294 \ub370\uc5d0\ub294 \uc5ec\ub7ec\uac00\uc9c0 \uc694\uc778\uc774 \uc788\uaca0\uc9c0\ub9cc, “\ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc774 \uc77c\uc744 \uc5bc\ub9c8\ub098 \uc798 \ud558\ub290\ub0d0”\ub294 \uace7 “\uc5bc\ub9c8\ub098 \ub9ce\uc740 \uc0ac\ub78c\ub4e4\uc774 \uba64\ubc84\uc27d\uc744 \uc720\uc9c0\ud558\ub290\ub0d0”\uc5d0\uc11c \uc5b4\ub290\uc815\ub3c4 \uce21\uc815\uc774 \uac00\ub2a5\ud558\ub2e4\uace0 \uc800\uc790\ub4e4\uc740 \uc774\uc57c\uae30\ud569\ub2c8\ub2e4. \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc774 \ub2e4\ub978 \ubd84\uc57c (NLP, CV, etc.) \ubcf4\ub2e4 \uc720\ub3c5 \uc5b4\ub824\uc6b4 \uc774\uc720\ub294 \uc0ac\ub78c\ub9c8\ub2e4 \ucde8\ud5a5\uacfc \ud589\ub3d9 \ud328\ud134\uc774 \ub9e4\uc6b0 \uc0c1\uc774\ud558\ub2e4\ub294 \ub370\uc5d0\uc11c \uc9d0\uc791\ud574\ubcfc \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uac00\ub839, \uc0ac\uc9c4\uc5d0\uc11c \uac1c\uc640 \uace0\uc591\uc774\ub97c \uad6c\ubd84\ud558\ub294 CV \ud0dc\uc2a4\ud06c\uc5d0\uc11c\ub294 \ub204\uad6c\uc5d0\uac8c\ub098 \ub3d9\uc77c\ud55c \uc815\ub2f5\uc774 \uc815\ud574\uc838 \uc788\uc9c0\ub9cc, \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc5d0\uc11c \uc5b4\ub5a4 \ucee8\ud150\uce20\ub97c \ucd94\ucc9c\ud560\uc9c0\ub294 \uc0ac\ub78c\ub9c8\ub2e4 \uc815\ub2f5\uc774 \ub2e4 \ub2e4\ub97c \uc218 \uc788\ub2e4\ub294 \uac83\uc785\ub2c8\ub2e4. <\/p>\n\n\n

\n
\"\"
\uadf8\ub9bc 1. Netflix \uccab \ud398\uc774\uc9c0\uc5d0\ub9cc \ud574\ub3c4 \uc5ec\ub7ec\uac00\uc9c0 \ucd94\ucc9c\uc2dc\uc2a4\ud15c \ud0dc\uc2a4\ud06c\uac00 \uc5f0\ub8e8\ub41c\ub2e4.<\/figcaption><\/figure>\n<\/div>\n\n\n

\uadf8\ub9bc 1\uc744 \ubcf4\uaca0\uc2b5\ub2c8\ub2e4. Netflix\uc758 \ud648\ud398\uc774\uc9c0\uc5d0\ub294 \uc0ac\uc2e4 \uc5ec\ub7ec\uac00\uc9c0 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\ub4e4\uc774 \uc801\uc6a9\ub418\uc5b4 \uc788\ub294\ub370\uc694, \uc608\ub97c \ub4e4\uc5b4 1\ubc88\uc758 \uacbd\uc6b0, \uc0ac\uc6a9\uc790\uac00 \uac00\uc7a5 \ubcf4\uace0\uc2f6\uc5b4\ud560 \ubc95\ud55c \ucee8\ud150\uce20\ub97c video \ud615\ud0dc\ub85c \ub744\uc6cc\uc8fc\ub294 \uacf5\uac04\uc774\uace0, 5\ubc88\uc758 \uacbd\uc6b0 Netflix\uc5d0\uc11c \ucd5c\uadfc \uc720\ud589\ud558\ub294 \ucee8\ud150\uce20\ub4e4\uc744 \ub098\uc5f4\ud574\uc8fc\ub294 \uacf5\uac04\uc774\uba70, 2\ubc88\uc740 \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998\uc774 \uc801\uc6a9\ub41c \uac80\uc0c9 \uc5d4\uc9c4\uc774 \uc790\ub9ac\ud558\uace0 \uc788\uc2b5\ub2c8\ub2e4. <\/p>\n\n\n\n

Modeling Approaches<\/h2>\n\n\n\n

\ub17c\ubb38\uc5d0\uc11c\ub294 Netflix\uc5d0 \uc801\uc6a9\ub41c \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998\ub4e4\uc744 \ud06c\uac8c \ub450 \uce74\ud14c\uace0\ub9ac\ub85c \ub098\ub204\uc5b4 \uc18c\uac1c\ud569\ub2c8\ub2e4. \uac01\uac01 Bag-of items \uce74\ud14c\uace0\ub9ac\uc640 Sequential model \uce74\ud14c\uace0\ub9ac\uc785\ub2c8\ub2e4. \uc8fc\ubaa9\ud560 \ub9cc\ud55c \ud2b9\uc9d5\uc740 \uc5ec\ub7ec heterogeneous \ud2b9\uc9d5\ub4e4\uc744 \uc0ac\uc6a9\ud560 \ub54c \ub525\ub7ec\ub2dd \uae30\ubc18\uc758 \uc54c\uace0\ub9ac\uc998\uc774 \uc88b\uc740 \uc131\ub2a5\uc744 \ubcf4\uc778 \ubc18\uba74, user-item interaction \ub370\uc774\ud130\ub9cc \uc0ac\uc6a9\ub420 \ub54c (\ub370\uc774\ud130\uac00 \ub2e8\uc21c\ud560 \ub54c)\uc5d0\ub294 \uac04\ub2e8\ud55c \uc804\ud1b5\uc801\uc778 \ubaa8\ub378\ub4e4\uc758 \uc131\ub2a5\uc774 \ub4a4\uccd0\uc9c0\uc9c0 \uc54a\uc558\ub2e4\ub294 \uc810\uc785\ub2c8\ub2e4. <\/p>\n\n\n\n

Bag-of-items Model<\/h3>\n\n\n\n

Bag-of-items\ub294 NLP\uc758 bag-of-words\ub97c \ub5a0\uc62c\ub9ac\uc2dc\uba74 \uc27d\uac8c \uc774\ud574\uac00 \ub418\uc2e4\ud150\ub370\uc694, \uc544\uc774\ud15c\uc758 \uc21c\uc11c\uc640 \uc0c1\uad00\uc5c6\uc774 \ubaa8\ub4e0 \uc544\uc774\ud15c (\ucee8\ud150\uce20)\ub4e4\uc744 \ubb49\ud145\uc774\ub85c \ud558\ub098\uc758 \uadf8\ub8f9\uc73c\ub85c \uc5ec\uae30\ub294 \ubc29\uc2dd\uc785\ub2c8\ub2e4. \ud6c8\ub828 \ub370\uc774\ud130\ub294 \uc0ac\uc6a9\uc790-\uc544\uc774\ud15c \ud589\ub82c\ub85c \ub098\ud0c0\ub0bc \uc218 \uc788\uc73c\uba70, Autoencoder, \ud589\ub82c \ubd84\ud574, \uc774\uc6c3 \uae30\ubc18 \ubc29\ubc95 \ub4f1 \ube44\uad50\uc801 \uc804\ud1b5\uc801\uc778 \uc54c\uace0\ub9ac\uc998\ub4e4\uc5d0 \uc801\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. <\/p>\n\n\n\n

\uc2e4\uc81c\ub85c Netflix Prize Competition (Bennet and Lanning 2007)\uc5d0\uc11c bag-of-items \ubc29\ubc95\ub860\ub4e4\uc774 \uc0c1\ub2f9\ud788 \ub9ce\uc774 \uc5f0\uad6c\ub418\uc5c8\uc73c\uba70 \ud765\ubbf8\ub86d\uac8c\ub3c4 \ub525\ub7ec\ub2dd \uae30\ubc18 \ubc29\ubc95\ub4e4\uacfc\uc758 \uc5f0\uad00\uc131\ub3c4 \ub9ce\uc558\ub2e4\uace0 \ud569\ub2c8\ub2e4. Bag-of-items \ubc29\uc2dd\uc774 Netflix \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc5d0\uc11c \ud6a8\uacfc\uc801\uc77c \uc218 \uc788\ub294 \uc774\uc720\ub294, (\uc2dc\ub9ac\uc988\ubb3c\uc774 \uc544\ub2cc \uc774\uc0c1) \ube44\ub514\uc624 \ub370\uc774\ud130\uac00 (NLP\uc758 text data \ub4f1\uacfc \ube44\uad50\ud588\uc744 \ub54c) \uc0c1\ub300\uc801\uc73c\ub85c \uc21c\uc11c\uc5d0 \ub35c \ubbfc\uac10\ud558\uae30 \ub54c\ubb38\uc774\ub77c\uace0 \uc800\uc790\ub4e4\uc740 \uc774\uc57c\uae30\ud558\uace0 \uc788\uc2b5\ub2c8\ub2e4. <\/p>\n\n\n\n

Sequential Models<\/h3>\n\n\n\n

Bag-of-items \ud615\ud0dc\uc758 \ub370\uc774\ud130\ub294 \uad6c\ucd95\ud558\uace0 \uc0ac\uc6a9\ud558\uae30\uc5d0 \ud3b8\ub9ac\ud558\uae34 \ud558\uc9c0\ub9cc, \uc544\ubb34\ub798\ub3c4 \uc0ac\uc6a9\uc790\uc758 \uc2dc\uccad \uc774\ub825 \uc21c\uc11c\ub97c \uace0\ub824\ud588\uc744 \ub54c \ubcf4\ub2e4 \ubcf5\uc7a1\ud55c \ud328\ud134\ub4e4\uc744 \ud559\uc2b5\ud558\uace0 \uc88b\uc740 \ucd94\ucc9c\uc744 \uc81c\uacf5\ud560 \uc218 \uc788\uaca0\uc8e0. \ub530\ub77c\uc11c \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc5d0\ub3c4 NLP\uc758 \uc21c\ucc28 \ubaa8\ub378\ub4e4\uc744 \uc801\uc6a9\ud55c \uc5f0\uad6c\uac00 \ub9ce\uc558\ub294\ub370, NLP\uc5d0\uc11c next word\ub97c \uc608\uce21\ud560 \ub54c, \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc5d0\uc11c\ub294 next item\uc744 \uc5d0\uce21\ud569\ub2c8\ub2e4.<\/p>\n\n\n\n

\uc2e4\uc81c Netflix \uc11c\ube44\uc2a4\uc5d0\uc11c \ub9ce\uc740 sub-task\ub4e4\uc740 bag-of-item \ubc29\ubc95\ub9cc\uc73c\ub85c\ub3c4 \ucda9\ubd84\ud558\uc9c0\ub9cc, \uc77c\ubd80 \ubcf5\uc7a1\ud55c sub-task\uc5d0\ub294 \uc21c\ucc28 \ubc29\ubc95\ub860\uc758 \uc774\uc810\uc774 \ub9ce\uc558\ub2e4\uace0 \ud569\ub2c8\ub2e4. \uc5f0\uad6c\ud300\uc5d0\uc11c\ub294 RNN \ubaa8\ub378\uc758 \uc77c\uc885\uc778 LSTM, GRU\ub97c \ube44\ub86f\ud574 Transformer \uc544\ud0a4\ud14d\ucc98 \uae30\ubc18\uc778 BERT \ub4f1 \ub2e4\uc591\ud55c \ubaa8\ub378\ub4e4\uc744 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc5d0 \uc801\uc6a9\uc2dc\ucf30\uace0, \ud2b9\ud788 Transformer\uc758 attention \ub9e4\ucee4\ub2c8\uc998\uc740 \ucd94\ucc9c\uc758 \uc815\ud655\ub3c4\ub97c \ud5a5\uc0c1\uc2dc\ucf30\uc744 \ubfd0\ub9cc \uc544\ub2c8\ub77c \ucd94\ucc9c \uacb0\uacfc\uc5d0 \uc5b4\ub290 \uc815\ub3c4\uc758 ‘\uc124\uba85\ub825’\uae4c\uc9c0 \uc81c\uacf5\ud574\uc8fc\ub294 \ud765\ubbf8\ub85c\uc6b4 \uacb0\uacfc\ub97c \uc5bb\uc5c8\ub2e4\uace0 \ud569\ub2c8\ub2e4.<\/p>\n\n\n

\n
\"\"
\uadf8\ub9bc 2. \ub17c\ubb38\uc5d0\uc11c \uc5b8\uae09\ub41c \ubaa8\ub378 \uac04 \uad00\uacc4\ub3c4<\/figcaption><\/figure>\n<\/div>\n\n\n

Adding Heterogeneous Features<\/h3>\n\n\n\n

\uc5f0\uad6c\uc9c4\ub4e4\uc740 \ud55c\ubc1c \uc55e\uc11c \ub098\uac00 “\uc804\ud1b5\uc801\uc778” \uc54c\uace0\ub9ac\uc998\ub4e4\uc774 \ud574\ub0b4\uae30 \uc5b4\ub824\uc6b4 \uc77c\ub4e4\uc774 \ubb34\uc5c7\uc774\uace0, \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998\uc774 \uc7a0\uc7ac\uc801\uc73c\ub85c \ud574\ub0bc \uc218 \uc788\ub294 \uc77c\uc774 \ubb34\uc5c7\uc774 \uc788\uc744\uc9c0 \uace0\ubbfc\ud574 \ubcf4\uc558\uc2b5\ub2c8\ub2e4. \uacb0\ub860\uc740 \ub525\ub7ec\ub2dd\uc740 \uc0ac\uc6a9\uc790\uc758 implicit\/explicit \ud53c\ub4dc\ubc31 \uc774\uc0c1\uc758 \ucd94\uac00\uc801\uc778 information\uc73c\ub85c\ubd80\ud130 \uc0c8\ub85c\uc6b4 \ubb54\uac00\ub97c \ub354 \ubf51\uc544\ub0b4\uc9c0 \uc54a\uc744\uae4c \ud558\ub294 \uac00\uc124\uc774\uc5c8\uace0, \ub9ce\uc740 \uc2e4\ud5d8\uc744 \ud1b5\ud574 \uc2e0\uc911\ud558\uac8c \uc120\uc815\ub41c heterogeneous feature\ub97c \ud65c\uc6a9\ud588\uc744 \ub54c \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \uc815\ud655\ub3c4\uac00 \ud06c\uac8c \ud5a5\uc0c1\ub418\uc5c8\ub2e4\uace0 \ud569\ub2c8\ub2e4. <\/p>\n\n\n\n

\uac00\ub839, time feature\uc744 \ud65c\uc6a9\ud588\uc744 \ub54c \ub525\ub7ec\ub2dd \ubaa8\ub378\uc740 seasonal information\uc774\ub098 cyclic element\ub97c \ucd94\uac00\uc801\uc73c\ub85c \ud559\uc2b5\ud560 \uc218 \uc788\uac8c \ub418\uc5c8\uc2b5\ub2c8\ub2e4. Time \ud2b9\uc131 \ud65c\uc6a9\uc758 \uc720\uc6a9\ud568\uc744 \uc54c\uac8c\ub418\uc790, \uc5f0\uad6c\uc9c4\uc740 time feature\uc744 \ub2e4\uc591\ud55c setting\uc73c\ub85c \uc124\uacc4\ud558\uc5ec baseline sequence prediction\uacfc \ube44\uad50\ud558\ub294 \uc2e4\ud5d8\uc744 \uc9c4\ud589\ud558\uc600\uc2b5\ub2c8\ub2e4.<\/p>\n\n\n

\n
\"\"
\uadf8\ub9bc 3. Time feature \ucd94\uac00\uc5d0 \ub530\ub978 \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc131\ub2a5 \ud5a5\uc0c1<\/figcaption><\/figure>\n<\/div>\n\n\n

\uadf8\ub9bc 3\uc5d0\uc11c \ubcf4\uc2dc\ub2e4\uc2dc\ud53c, offline evaluation \uacb0\uacfc \ud2b9\ud788 continuous-time feature\uc744 \ud65c\uc6a9\ud588\uc744 \ub54c baseline\uc5d0\uc11c \ud06c\uac8c \ud5a5\uc0c1\ud55c \uc131\ub2a5\uc744 \uad00\ucc30\ud560 \uc218 \uc788\uc5c8\uc2b5\ub2c8\ub2e4.<\/p>\n\n\n\n

Discussions & Others<\/h2>\n\n\n\n
    \n
  1. \ub370\uc774\ud130\uc758 \uc911\uc694\uc131: \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998 \uadf8 \uc790\uccb4 \ub9cc\ud07c\uc774\ub098 \uc911\uc694\ud55c \uac83\uc740 \ub370\uc774\ud130\uc785\ub2c8\ub2e4. \ub370\uc774\ud130\uac00 user-item interaction \uc815\ubcf4\ub9cc \uac00\uc9c0\uace0 \uc788\ub294\uc9c0, user feature, item feature\uacfc \uac19\uc740 side information, \ub610\ub294 contextual information\uae4c\uc9c0 \ud3ec\ud568\ud558\uace0 \uc788\ub294\uc9c0\uc5d0 \ub530\ub77c \uc0ac\uc6a9\ud558\uace0 \uac1c\ubc1c\ud560 \uc218 \uc788\ub294 \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998 \ubc94\uc704\uac00 \ud06c\uac8c \ub2ec\ub77c\uc9c8 \uc218 \uc788\uc2b5\ub2c8\ub2e4. <\/li>\n\n\n\n
  2. Bias \ubb38\uc81c: \ucee8\ud150\uce20\ub97c \ub744\uc6cc\uc8fc\ub294 \uc704\uce58\uc5d0\uc11c \ubc1c\uc0dd\ud558\ub294 presentation\/position bias\ub97c \ud3ec\ud568\ud55c \ub2e4\uc591\ud55c bias\ub4e4\uc740 \ub525\ub7ec\ub2dd \uae30\ubc18 \ucd94\ucc9c\uc2dc\uc2a4\ud15c \ud559\uc2b5\uc744 \uc5b4\ub835\uac8c \ub9cc\ub4dc\ub294 \uc694\uc18c\uac00 \ub418\uae30\ub3c4 \ud569\ub2c8\ub2e4.<\/li>\n\n\n\n
  3. Offline metrics\uc5d0\uc11c evlauation\uc744 \ud1b5\uacfc\ud55c \ubc29\ubc95\ub4e4\uc744 \uac00\uc9c0\uace0 A\/B \ud14c\uc2a4\ud2b8\ub97c \ud588\uc744 \ub54c \uc5b8\uc81c\ub098 \uac19\uc740 \uc88b\uc740 \uacb0\uacfc\uac00 \ub098\uc624\uc9c4 \uc54a\b\uae30\uc5d0, \uc0c8\ub85c\uc6b4 (online \uc138\ud305\uc5d0\uc11c\uc758 \uacb0\uacfc\uc640 \ube44\uc2b7\ud560\ub9cc\ud55c) offline metrics\uc744 \uace0\uc548\ud574\ub0b4\ub294 \uac83 \ub610\ud55c \ub9e4\uc6b0 \uc911\uc694\ud558\ub2e4\uace0 \ud569\ub2c8\ub2e4.<\/li>\n<\/ol>\n\n\n\n

    \ub9c8\uce58\uba70<\/h2>\n\n\n\n

    \ucd94\ucc9c\uc2dc\uc2a4\ud15c \ubd84\uc57c\uc758 \ubc1c\uc804\uc5d0 \uc5b4\ucc0c \ubcf4\uba74 \uac00\uc7a5 \ud06c\uac8c \uae30\uc5ec\ud55c \uae30\uc5c5\uc778 Netflix \uc5f0\uad6c\ud300\uc758 struggle\uacfc achievement\ub97c \uac04\uc811\uc801\uc73c\ub85c \uacbd\ud5d8\ud560 \uc218 \uc788\uc5b4\uc11c \uc7ac\ubbf8\uc788\ub294 \ub17c\ubb38\uc774\uc5c8\uc2b5\ub2c8\ub2e4. \ud14c\ud06c\ub2c8\uceec\ud55c \uae4a\uc740 \ub0b4\uc6a9\uc774 \ub9ce\uc9c0\ub294 \uc54a\uc544\uc11c, \ud3c9\uc18c AI\uc5d0 \uad00\uc2ec\uc788\ub294 \ube44\uc804\uacf5\uc790 \ubd84\ub4e4\ub3c4 \ucda9\ubd84\ud788 \uc5bb\uc5b4\uac08 \ub9cc\ud55c \ub0b4\uc6a9\uc774 \ub9ce\uc740 \uc544\ud2f0\ud074\uc774 \uc544\ub2cc\uac00 \uc2f6\uc2b5\ub2c8\ub2e4. \uc55e\uc73c\ub85c \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc5f0\uad6c\uc758 \ud328\ub7ec\ub2e4\uc784\uc774 \uc5b4\ub5a4 \ubcc0\ud654\ub97c \uacaa\uc744\uc9c0 \uad81\uae08\ud558\uae30\ub3c4 \ud569\ub2c8\ub2e4. Netflix\uac00 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc73c\ub85c \uba64\ubc84\uc27d \uac00\uc785\uc744 \ub298\ub9ac\ub294 \ub370 \ub354 \ub3c4\uc6c0 \uc904 space\uac00 \uc5bc\ub9c8\ub098 \ub0a8\uc544\uc788\uc744 \uc9c0 \uad81\uae08\ud558\uae30\ub3c4 \ud558\uace0\uc694.<\/p>\n\n\n\n

    <\/p>\n","protected":false},"excerpt":{"rendered":"

    Netflix\uc758 research scientist \uc800\uc790\ub4e4\uc774 \uc791\uc131\ud55c \ub17c\ubb38 “Deep learning for recommender systems: A Netflix case study” \ub9ac\ubdf0\uc785\ub2c8\ub2e4. \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998 \uc5ed\uc0ac\uc5d0 \ud070 \ud55c \ud68d\uc744 \uadf8\uc740 \uae30\uc5c5\uc778 Netflix\uc758 \uc5f0\uad6c\ud300\uc5d0\uc11c \ud37c\ube14\ub9ac\uc2dc\ud55c \ub17c\ubb38\uc778 \ub9cc\ud07c, \uadf8\ub4e4\uc758 struggle\uacfc achievement\ub97c \uac04\uc811\uc801\uc73c\ub85c \ub290\ub07c\uba70 \uac00\ubccd\uac8c(?) \uc77d\uc5b4\ubcf4\uae30\uc5d0 \uc88b\uc740 \uc544\ud2f0\ud074\uc774\uc5c8\uc2b5\ub2c8\ub2e4. Introduction \uc9c0\ub09c \uc218 \ub144\uac04 […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"nf_dc_page":"","_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[64],"tags":[67,66,65],"class_list":["post-447","post","type-post","status-publish","format-standard","hentry","category-recommender-system","tag-netflix","tag-paper-review","tag-recsys"],"yoast_head":"\n\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0) - Jaein's Lab<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/jaeinlab.com\/?p=447\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0) - Jaein's Lab\" \/>\n<meta property=\"og:description\" content=\"Netflix\uc758 research scientist \uc800\uc790\ub4e4\uc774 \uc791\uc131\ud55c \ub17c\ubb38 “Deep learning for recommender systems: A Netflix case study” \ub9ac\ubdf0\uc785\ub2c8\ub2e4. \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998 \uc5ed\uc0ac\uc5d0 \ud070 \ud55c \ud68d\uc744 \uadf8\uc740 \uae30\uc5c5\uc778 Netflix\uc758 \uc5f0\uad6c\ud300\uc5d0\uc11c \ud37c\ube14\ub9ac\uc2dc\ud55c \ub17c\ubb38\uc778 \ub9cc\ud07c, \uadf8\ub4e4\uc758 struggle\uacfc achievement\ub97c \uac04\uc811\uc801\uc73c\ub85c \ub290\ub07c\uba70 \uac00\ubccd\uac8c(?) \uc77d\uc5b4\ubcf4\uae30\uc5d0 \uc88b\uc740 \uc544\ud2f0\ud074\uc774\uc5c8\uc2b5\ub2c8\ub2e4. Introduction \uc9c0\ub09c \uc218 \ub144\uac04 […]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/jaeinlab.com\/?p=447\" \/>\n<meta property=\"og:site_name\" content=\"Jaein's Lab\" \/>\n<meta property=\"article:published_time\" content=\"2024-01-11T11:18:16+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2024-01-11T11:21:11+00:00\" \/>\n<meta name=\"author\" content=\"Jaein\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Jaein\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/jaeinlab.com\\\/?p=447#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/jaeinlab.com\\\/?p=447\"},\"author\":{\"name\":\"Jaein\",\"@id\":\"http:\\\/\\\/jaeinlab.com\\\/#\\\/schema\\\/person\\\/70edb94493e55ab8edbafc48bdf56883\"},\"headline\":\"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0)\",\"datePublished\":\"2024-01-11T11:18:16+00:00\",\"dateModified\":\"2024-01-11T11:21:11+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/jaeinlab.com\\\/?p=447\"},\"wordCount\":152,\"commentCount\":0,\"keywords\":[\"Netflix\",\"paper review\",\"RecSys\"],\"articleSection\":[\"Recommender System\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/jaeinlab.com\\\/?p=447#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/jaeinlab.com\\\/?p=447\",\"url\":\"https:\\\/\\\/jaeinlab.com\\\/?p=447\",\"name\":\"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0) - Jaein's Lab\",\"isPartOf\":{\"@id\":\"http:\\\/\\\/jaeinlab.com\\\/#website\"},\"datePublished\":\"2024-01-11T11:18:16+00:00\",\"dateModified\":\"2024-01-11T11:21:11+00:00\",\"author\":{\"@id\":\"http:\\\/\\\/jaeinlab.com\\\/#\\\/schema\\\/person\\\/70edb94493e55ab8edbafc48bdf56883\"},\"breadcrumb\":{\"@id\":\"https:\\\/\\\/jaeinlab.com\\\/?p=447#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/jaeinlab.com\\\/?p=447\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/jaeinlab.com\\\/?p=447#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"http:\\\/\\\/jaeinlab.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0)\"}]},{\"@type\":\"WebSite\",\"@id\":\"http:\\\/\\\/jaeinlab.com\\\/#website\",\"url\":\"http:\\\/\\\/jaeinlab.com\\\/\",\"name\":\"Jaein's Lab\",\"description\":\"Natural Language Processing, Knowledge Graph, and others!\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"http:\\\/\\\/jaeinlab.com\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"http:\\\/\\\/jaeinlab.com\\\/#\\\/schema\\\/person\\\/70edb94493e55ab8edbafc48bdf56883\",\"name\":\"Jaein\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/6a23798f753744134b0f342a342b0fb0d4035a7313d8b7c54faf930978eaa942?s=96&d=wp_user_avatar&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/6a23798f753744134b0f342a342b0fb0d4035a7313d8b7c54faf930978eaa942?s=96&d=wp_user_avatar&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/6a23798f753744134b0f342a342b0fb0d4035a7313d8b7c54faf930978eaa942?s=96&d=wp_user_avatar&r=g\",\"caption\":\"Jaein\"},\"description\":\"I'm a Computer Science PhD candidate from South Korea, currently based in Beijing, China. My academic journey is dedicated to pioneering research in Natural Language Processing and Knowledge Graphs, striving to break new ground in these dynamic fields.\",\"sameAs\":[\"http:\\\/\\\/jaeinlab.com\",\"https:\\\/\\\/www.linkedin.com\\\/in\\\/jaeinlab\\\/\"],\"url\":\"https:\\\/\\\/jaeinlab.com\\\/?author=1\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0) - Jaein's Lab","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/jaeinlab.com\/?p=447","og_locale":"en_US","og_type":"article","og_title":"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0) - Jaein's Lab","og_description":"Netflix\uc758 research scientist \uc800\uc790\ub4e4\uc774 \uc791\uc131\ud55c \ub17c\ubb38 “Deep learning for recommender systems: A Netflix case study” \ub9ac\ubdf0\uc785\ub2c8\ub2e4. \ucd94\ucc9c\uc2dc\uc2a4\ud15c \uc54c\uace0\ub9ac\uc998 \uc5ed\uc0ac\uc5d0 \ud070 \ud55c \ud68d\uc744 \uadf8\uc740 \uae30\uc5c5\uc778 Netflix\uc758 \uc5f0\uad6c\ud300\uc5d0\uc11c \ud37c\ube14\ub9ac\uc2dc\ud55c \ub17c\ubb38\uc778 \ub9cc\ud07c, \uadf8\ub4e4\uc758 struggle\uacfc achievement\ub97c \uac04\uc811\uc801\uc73c\ub85c \ub290\ub07c\uba70 \uac00\ubccd\uac8c(?) \uc77d\uc5b4\ubcf4\uae30\uc5d0 \uc88b\uc740 \uc544\ud2f0\ud074\uc774\uc5c8\uc2b5\ub2c8\ub2e4. Introduction \uc9c0\ub09c \uc218 \ub144\uac04 […]","og_url":"https:\/\/jaeinlab.com\/?p=447","og_site_name":"Jaein's Lab","article_published_time":"2024-01-11T11:18:16+00:00","article_modified_time":"2024-01-11T11:21:11+00:00","author":"Jaein","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Jaein","Est. reading time":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/jaeinlab.com\/?p=447#article","isPartOf":{"@id":"https:\/\/jaeinlab.com\/?p=447"},"author":{"name":"Jaein","@id":"http:\/\/jaeinlab.com\/#\/schema\/person\/70edb94493e55ab8edbafc48bdf56883"},"headline":"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0)","datePublished":"2024-01-11T11:18:16+00:00","dateModified":"2024-01-11T11:21:11+00:00","mainEntityOfPage":{"@id":"https:\/\/jaeinlab.com\/?p=447"},"wordCount":152,"commentCount":0,"keywords":["Netflix","paper review","RecSys"],"articleSection":["Recommender System"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/jaeinlab.com\/?p=447#respond"]}]},{"@type":"WebPage","@id":"https:\/\/jaeinlab.com\/?p=447","url":"https:\/\/jaeinlab.com\/?p=447","name":"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0) - Jaein's Lab","isPartOf":{"@id":"http:\/\/jaeinlab.com\/#website"},"datePublished":"2024-01-11T11:18:16+00:00","dateModified":"2024-01-11T11:21:11+00:00","author":{"@id":"http:\/\/jaeinlab.com\/#\/schema\/person\/70edb94493e55ab8edbafc48bdf56883"},"breadcrumb":{"@id":"https:\/\/jaeinlab.com\/?p=447#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/jaeinlab.com\/?p=447"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/jaeinlab.com\/?p=447#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"http:\/\/jaeinlab.com\/"},{"@type":"ListItem","position":2,"name":"\ub137\ud50c\ub9ad\uc2a4 \ucd94\ucc9c\uc2dc\uc2a4\ud15c\uc758 \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ubc1c\uc804\uc0ac (\ud398\uc774\ud37c \ub9ac\ubdf0)"}]},{"@type":"WebSite","@id":"http:\/\/jaeinlab.com\/#website","url":"http:\/\/jaeinlab.com\/","name":"Jaein's Lab","description":"Natural Language Processing, Knowledge Graph, and others!","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"http:\/\/jaeinlab.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"http:\/\/jaeinlab.com\/#\/schema\/person\/70edb94493e55ab8edbafc48bdf56883","name":"Jaein","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/6a23798f753744134b0f342a342b0fb0d4035a7313d8b7c54faf930978eaa942?s=96&d=wp_user_avatar&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/6a23798f753744134b0f342a342b0fb0d4035a7313d8b7c54faf930978eaa942?s=96&d=wp_user_avatar&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/6a23798f753744134b0f342a342b0fb0d4035a7313d8b7c54faf930978eaa942?s=96&d=wp_user_avatar&r=g","caption":"Jaein"},"description":"I'm a Computer Science PhD candidate from South Korea, currently based in Beijing, China. My academic journey is dedicated to pioneering research in Natural Language Processing and Knowledge Graphs, striving to break new ground in these dynamic fields.","sameAs":["http:\/\/jaeinlab.com","https:\/\/www.linkedin.com\/in\/jaeinlab\/"],"url":"https:\/\/jaeinlab.com\/?author=1"}]}},"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/jaeinlab.com\/index.php?rest_route=\/wp\/v2\/posts\/447","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/jaeinlab.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/jaeinlab.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/jaeinlab.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/jaeinlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=447"}],"version-history":[{"count":5,"href":"https:\/\/jaeinlab.com\/index.php?rest_route=\/wp\/v2\/posts\/447\/revisions"}],"predecessor-version":[{"id":459,"href":"https:\/\/jaeinlab.com\/index.php?rest_route=\/wp\/v2\/posts\/447\/revisions\/459"}],"wp:attachment":[{"href":"https:\/\/jaeinlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=447"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/jaeinlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=447"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/jaeinlab.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=447"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}