{"id":2684,"date":"2021-01-29T16:35:07","date_gmt":"2021-01-29T16:35:07","guid":{"rendered":"https:\/\/thenextweb.com\/?p=1336629"},"modified":"2021-01-29T16:35:07","modified_gmt":"2021-01-29T16:35:07","slug":"mits-new-liquid-neural-network-learns-on-the-job-so-robots-can-adapt-to-changing-conditions","status":"publish","type":"post","link":"https:\/\/www.londonchiropracter.com\/?p=2684","title":{"rendered":"MIT\u2019s new \u2018liquid\u2019 neural network learns on the job \u2014 so robots can adapt to changing conditions"},"content":{"rendered":"\n<div><img decoding=\"async\" src=\"https:\/\/img-cdn.tnwcdn.com\/image\/neural?filter_last=1&amp;fit=1280%2C640&amp;url=https%3A%2F%2Fcdn0.tnwcdn.com%2Fwp-content%2Fblogs.dir%2F1%2Ffiles%2F2021%2F01%2FUntitled-design-2021-01-29T145833.374.png&amp;signature=493814af6f23ef16fc1e977580199ff0\" class=\"ff-og-image-inserted\"><\/div>\n<p>MIT researchers have invented an adaptive \u201cliquid\u201d neural network&nbsp;that could improve decision-making in self-driving cars and medical diagnosis.<\/p>\n<p>The algorithm adjusts to changes experienced by real-world systems by changing their underlying equations as they receive new data.<\/p>\n<p>\u201cThis is a way forward for the future of robot control, natural language processing, video processing \u2014 any form of time series data processing,\u201d <a href=\"https:\/\/news.mit.edu\/2021\/machine-learning-adapts-0128\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">said<\/a> Ramin Hasani, the study paper\u2019s&nbsp;lead author. \u201cThe potential is really significant.\u201d<\/p>\n<p><em>[Read: <a href=\"https:\/\/thenextweb.com\/artificial-intelligence\/2021\/01\/21\/how-this-company-leveraged-ai-to-become-the-netflix-of-finland\/\">How this company leveraged AI to become the Netflix of Finland<\/a>]<\/em><\/p>\n<p>Hasani said the system is inspired by a tiny worm \u2014 the C. elegans:<\/p>\n<blockquote readability=\"5\">\n<p>It only has 302 neurons in its nervous system yet it can generate unexpectedly complex dynamics.<\/p>\n<\/blockquote>\n<p>The&nbsp;code was influenced by the way the C. elegans\u2019 neurons activate and communicate with each other through electrical impulses.<\/p>\n<p>Hasani structed his neural network so that the parameters can change over time based on the results of a nested set of differential equations.<\/p>\n<p>This allows it to continue learning after the training phase, making it more resilient to unexpected situations, like heavy rain covering a camera on a self-driving car.<\/p>\n<h2>Future plans<\/h2>\n<p>The liquid network\u2019s small number of highly expressive neurons also makes it easier to interpret its decisions.<\/p>\n<p>\u201cJust [by] changing the representation of a neuron, you can really explore some degrees of complexity you couldn\u2019t explore otherwise,\u201d said Hasani.<\/p>\n<p>In tests, the network performed promisingly in predicting future values in datasets, ranging&nbsp;from atmospheric chemistry to traffic patterns.<\/p>\n<p>Its small size also significantly reduced the computing costs.<\/p>\n<p>Hasani said he now wants to prepare the system for real-world applications:<\/p>\n<blockquote readability=\"7\">\n<p>We have a provably more expressive neural network that is inspired by nature. But this is just the beginning of the process. The obvious question is how do you extend this? We think this kind of network could be a key element of future intelligence systems.<\/p>\n<\/blockquote>\n<p>You can read the study paper on the pre-print server <a href=\"https:\/\/arxiv.org\/pdf\/2006.04439.pdf\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">arXiv<\/a>.<\/p>\n<p class=\"c-post-pubDate\"> Published January 29, 2021 \u2014 16:35 UTC <\/p>\n<p> <a href=\"https:\/\/thenextweb.com\/neural\/2021\/01\/29\/mits-new-liquid-neural-network-learns-on-the-job-so-robots-can-adapt-to-changing-conditions\/\">Source<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>MIT researchers have invented an adaptive \u201cliquid\u201d neural network&nbsp;that could improve decision-making in self-driving cars and medical diagnosis. The algorithm adjusts to changes experienced by real-world systems by changing their underlying equations&#8230;<\/p>\n","protected":false},"author":1,"featured_media":2685,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[],"_links":{"self":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/posts\/2684"}],"collection":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=2684"}],"version-history":[{"count":0,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/posts\/2684\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/media\/2685"}],"wp:attachment":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2684"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2684"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2684"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}