{"id":125567,"date":"2021-07-27T17:23:57","date_gmt":"2021-07-28T00:23:57","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2021\/07\/deepmind-generally-capable-agents-emerge-from-open-ended-play"},"modified":"2021-07-27T17:23:57","modified_gmt":"2021-07-28T00:23:57","slug":"deepmind-generally-capable-agents-emerge-from-open-ended-play","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2021\/07\/deepmind-generally-capable-agents-emerge-from-open-ended-play","title":{"rendered":"DeepMind: Generally capable agents emerge from open-ended play"},"content":{"rendered":"<p><\/p>\n<p><iframe style=\"display: block; margin: 0 auto; width: 100%; aspect-ratio: 4\/3; object-fit: contain;\" src=\"https:\/\/www.youtube.com\/embed\/lTmL7jwFfdw?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; encrypted-media; gyroscope;\n   picture-in-picture\" allowfullscreen><\/iframe><\/p>\n<p>I\u2019ve been suggesting for a long time to drop these Ai\u2019s into open world games.<\/p>\n<hr>\n<p>EDIT: Also see paper and results compilation video!<\/p>\n<p>Today, we published \u201cOpen-Ended Learning Leads to Generally Capable Agents,\u201d a preprint detailing our first steps to train an agent capable of playing many different games without needing human interaction data. \u2026 The result is an agent with the ability to succeed at a wide spectrum of tasks \u2014 from simple object-finding problems to complex games like hide and seek and capture the flag, which were not encountered during training. We find the agent exhibits general, heuristic behaviours such as experimentation, behaviours that are widely applicable to many tasks rather than specialised to an individual task.<\/p>\n<p>The neural network architecture we use provides an attention mechanism over the agent\u2019s internal recurrent state \u2014 helping guide the agent\u2019s attention with estimates of subgoals unique to the game the agent is playing.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>I\u2019ve been suggesting for a long time to drop these Ai\u2019s into open world games. EDIT: Also see paper and results compilation video! Today, we published \u201cOpen-Ended Learning Leads to Generally Capable Agents,\u201d a preprint detailing our first steps to train an agent capable of playing many different games without needing human interaction data. \u2026 [\u2026]<\/p>\n","protected":false},"author":359,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1509,6],"tags":[],"class_list":["post-125567","post","type-post","status-publish","format-standard","hentry","category-entertainment","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/125567","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/users\/359"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=125567"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/125567\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=125567"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=125567"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=125567"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}