{"id":332238,"date":"2017-02-02T06:00:03","date_gmt":"2017-02-01T22:00:03","guid":{"rendered":"http:\/\/techcrunch.com\/?p=1446459"},"modified":"2017-02-02T06:00:03","modified_gmt":"2017-02-01T22:00:03","slug":"how-facebook-plans-to-evaluate-its-quest-for-generalized-artificial-intelligence","status":"publish","type":"post","link":"https:\/\/people.utm.my\/asmawisham\/how-facebook-plans-to-evaluate-its-quest-for-generalized-artificial-intelligence\/","title":{"rendered":"How Facebook plans to evaluate its quest for generalized artificial intelligence"},"content":{"rendered":"<p><img decoding=\"async\" src=\"https:\/\/tctechcrunch2011.files.wordpress.com\/2017\/02\/gettyimages-548761131.jpg?w=738\" class=\"\"\/><\/p>\n<p>One of the biggest\u00a0misconceptions about artificial intelligence is the belief that today\u2019s AIs\u00a0possess generalized intelligence. We are really good at\u00a0leveraging large data sets to accomplish\u00a0specific tasks, but fall flat at replicating the breadth of human intelligence.<\/p>\n<p>If we\u2019re going to move toward generalized intelligence, Facebook wants to make sure we know how to evaluate progress. <a target=\"_blank\" href=\"https:\/\/arxiv.org\/pdf\/1701.08954v1.pdf\">In a paper,<\/a> Facebook\u2019s AI Research (<a target=\"_blank\" href=\"https:\/\/research.fb.com\/category\/facebook-ai-research-fair\/\">FAIR<\/a>) lab outlines just that as part of its <a target=\"_blank\" href=\"https:\/\/github.com\/facebookresearch\/CommAI-env\/\">CommAI framework.\u00a0\u00a0<\/a><\/p>\n<p>First, the\u00a0team\u00a0argues that a proper generalized AI should be able to fluently communicate\u00a0in natural language with humans.\u00a0Researchers in the field of natural language processing have been working on the problem of human computer interaction through language for some time now, but we have a long way to go before our computational systems can rival humans.<\/p>\n<p>This is\u00a0obvious if you\u2019ve spent any time talking to the latest, greatest conversational AI systems on the market, like Apple\u2019s Siri and Amazon\u2019s Alexa. We will need our systems to be able to communicate and learn through language effectively, even when lacking context and discussing things in undefined terms.<\/p>\n<p>Moreover, such a system should be capable of learning new skills easily. Facebook calls this skill set \u201clearning to learn.\u201d Today\u2019s machine learning models can be trained on data and used for classifying defined objects. We can use transfer learning to quickly adapt a model to\u00a0accomplish the same task on new data, but\u00a0our machines can\u2019t\u00a0teach themselves completely new skills without heavy intervention from developers.<\/p>\n<div class=\"page section layoutArea column\" title=\"Page 1\" readability=\"39\">\n<p>\u201cIt\u2019s generally agreed that, in order to generalize across tasks, a program should be capable of compositional learning, that is, of storing and re-combining solutions to sub-problems across tasks,\u201d the team adds.<\/p>\n<\/div>\n<p>In a nod to the growing sub-field of reinforcement learning, Facebook\u00a0also notes that a generalized AI should resemble a human\u2019s ability to master new tasks with decreasing explicit rewards. And with respect to interface, modern systems should be able to take in new information and express themselves\u00a0in a variety of ways that\u00a0fit with the demands of diverse situations.<\/p>\n<p>Facebook considers these capabilities to be more of a prerequisite to generalized AI than a true <a target=\"_blank\" href=\"https:\/\/en.wikipedia.org\/wiki\/Turing_test\">Turing test<\/a>. The original Turing test was created by Alan Turing in 1950. It\u2019s generally understood to be a means of evaluating machine intelligence with respect to human intelligence.<\/p>\n<p>But as the field of AI has matured, the test has lost much\u00a0of its relevance. Facebook provides a nice alternative way to think about the important requirements of a modern generalized AI that should be less of a research distraction than the more rigid Turing test.<\/p>\n<p>The FAIR team of Marco Baroni, Armand Joulin, Allan Jabri, Germ\u00e1n Kruszewski, Angeliki Lazaridou, Klemen Simonic and Tomas Mikolov also developed an open source platform for testing and training AI systems. Similar to OpenAI\u2019s Gym, <a target=\"_blank\" href=\"https:\/\/github.com\/facebookresearch\/CommAI-env\/\">the\u00a0CommAI-env tool<\/a> focuses on more incremental tasks that build on previous accomplishments.\u00a0It\u2019s also designed to push models to their limits by emphasizing task variation.<\/p>\n<p><small>Featured Image: GraphicaArtis\/Getty Images<\/small><\/p>\n","protected":false},"excerpt":{"rendered":"<p><img loading=\"lazy\" decoding=\"async\" width=\"680\" height=\"518\" src=\"https:\/\/tctechcrunch2011.files.wordpress.com\/2017\/02\/gettyimages-548761131.jpg?w=680\" class=\"attachment-large size-large wp-post-image\" alt=\"Man With Circuit Board Brain\" srcset=\"https:\/\/tctechcrunch2011.files.wordpress.com\/2017\/02\/gettyimages-548761131.jpg?w=680 680w, https:\/\/tctechcrunch2011.files.wordpress.com\/2017\/02\/gettyimages-548761131.jpg?w=1360 1360w, https:\/\/tctechcrunch2011.files.wordpress.com\/2017\/02\/gettyimages-548761131.jpg?w=150 150w, https:\/\/tctechcrunch2011.files.wordpress.com\/2017\/02\/gettyimages-548761131.jpg?w=300 300w, https:\/\/tctechcrunch2011.files.wordpress.com\/2017\/02\/gettyimages-548761131.jpg?w=768 768w\" sizes=\"(max-width: 680px) 100vw, 680px\" \/>&nbsp;One of the biggest\u00a0misconceptions about artificial intelligence is the belief that today&#8217;s AIs\u00a0possess generalized intelligence. We are really good at\u00a0leveraging large data sets to accomplish\u00a0specific tasks, but fall flat at replicating the breadth of human intelligence. If we&#8217;re going to move toward generalized intelligence, Facebook wants to make sure we know how to&hellip; <a href=\"https:\/\/techcrunch.com\/2017\/02\/01\/how-facebook-plans-to-evaluate-its-quest-for-generalized-artificial-intelligence\/?ncid=rss\">Read More<\/a><\/p>\n<div class=\"feedflare\">\n<a href=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?a=UjxxLj6lnhk:ouVp4XPfRjQ:2mJPEYqXBVI\"><img decoding=\"async\" src=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?d=2mJPEYqXBVI\" border=\"0\"><\/img><\/a> <a href=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?a=UjxxLj6lnhk:ouVp4XPfRjQ:7Q72WNTAKBA\"><img decoding=\"async\" src=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?d=7Q72WNTAKBA\" border=\"0\"><\/img><\/a> <a href=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?a=UjxxLj6lnhk:ouVp4XPfRjQ:yIl2AUoC8zA\"><img decoding=\"async\" src=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?d=yIl2AUoC8zA\" border=\"0\"><\/img><\/a> <a href=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?a=UjxxLj6lnhk:ouVp4XPfRjQ:-BTjWOF_DHI\"><img decoding=\"async\" src=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?i=UjxxLj6lnhk:ouVp4XPfRjQ:-BTjWOF_DHI\" border=\"0\"><\/img><\/a> <a href=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?a=UjxxLj6lnhk:ouVp4XPfRjQ:D7DqB2pKExk\"><img decoding=\"async\" src=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?i=UjxxLj6lnhk:ouVp4XPfRjQ:D7DqB2pKExk\" border=\"0\"><\/img><\/a> <a href=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?a=UjxxLj6lnhk:ouVp4XPfRjQ:qj6IDK7rITs\"><img decoding=\"async\" src=\"http:\/\/feeds.feedburner.com\/~ff\/Techcrunch?d=qj6IDK7rITs\" border=\"0\"><\/img><\/a>\n<\/div>\n<p><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/feeds.feedburner.com\/~r\/Techcrunch\/~4\/UjxxLj6lnhk\" height=\"1\" width=\"1\" alt=\"\"\/> <\/p>\n","protected":false},"author":5817,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_et_pb_use_builder":"","_et_pb_old_content":"","_et_gb_content_width":"","footnotes":""},"categories":[25],"tags":[59,66,26],"class_list":["post-332238","post","type-post","status-publish","format-standard","hentry","category-technology","tag-media","tag-techcrunch","tag-technology"],"_links":{"self":[{"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/posts\/332238","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/users\/5817"}],"replies":[{"embeddable":true,"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/comments?post=332238"}],"version-history":[{"count":0,"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/posts\/332238\/revisions"}],"wp:attachment":[{"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/media?parent=332238"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/categories?post=332238"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/people.utm.my\/asmawisham\/wp-json\/wp\/v2\/tags?post=332238"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}