{"id":89266,"date":"2019-04-05T16:22:29","date_gmt":"2019-04-05T23:22:29","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2019\/04\/using-ai-to-make-better-ai"},"modified":"2019-04-05T16:22:29","modified_gmt":"2019-04-05T23:22:29","slug":"using-ai-to-make-better-ai","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2019\/04\/using-ai-to-make-better-ai","title":{"rendered":"Using AI to Make Better AI"},"content":{"rendered":"<p><a class=\"aligncenter blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/using-ai-to-make-better-ai2.jpg\"><\/a><\/p>\n<p>Next month, however, a team of <a href=\"http:\/\/www.mit.edu\/\">MIT<\/a> researchers will be presenting a so-called <a href=\"http:\/\/arxiv.org\/abs\/1812.00332\">\u201cProxyless<\/a> <a href=\"https:\/\/arxiv.org\/abs\/1812.00332\" target=\"_blank\">neural architecture search\u201d algorithm<\/a> that can speed up the AI-optimized AI design process by 240 times or more. That would put faster and more accurate AI within practical reach for a broad class of image recognition algorithms and other related applications.<\/p>\n<p>\u201cThere are all kinds of tradeoffs between model size, inference latency, accuracy, and model capacity,\u201d says <a href=\"https:\/\/songhan.mit.edu\/\">Song Han<\/a>, assistant professor of electrical engineering and computer science at MIT. Han adds that:<\/p>\n<p>\u201c[These] all add up to a giant design space. Previously people had designed neural networks based on heuristics. Neural architecture search tried to free this labor intensive, human heuristic-based exploration [by turning it] into a learning-based, AI-based design space exploration. Just like AI can [learn to] <a href=\"https:\/\/en.wikipedia.org\/wiki\/AlphaGo_versus_Lee_Sedol\" target=\"_blank\">play a Go game<\/a>, AI can [learn how to] design a neural network.\u201d<\/p>\n<p><!-- Link: <a href=\"https:\/\/spectrum.ieee.org\/tech-talk\/computing\/networks\/using-ai-to-make-better-ai\">https:\/\/spectrum.ieee.org\/tech-talk\/computing\/networks\/using...-better-ai<\/a> --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Next month, however, a team of MIT researchers will be presenting a so-called \u201cProxyless neural architecture search\u201d algorithm that can speed up the AI-optimized AI design process by 240 times or more. That would put faster and more accurate AI within practical reach for a broad class of image recognition algorithms and other related applications. [\u2026]<\/p>\n","protected":false},"author":513,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[41,6,1514],"tags":[],"class_list":["post-89266","post","type-post","status-publish","format-standard","hentry","category-information-science","category-robotics-ai","category-space-travel"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/89266","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\/513"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=89266"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/89266\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=89266"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=89266"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=89266"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}