{"id":48090,"date":"2019-12-13T05:53:44","date_gmt":"2019-12-13T04:53:44","guid":{"rendered":"https:\/\/www.rudebaguette.com\/?p=48090"},"modified":"2019-12-13T05:53:54","modified_gmt":"2019-12-13T04:53:54","slug":"waymo-acquires-the-uk-firm-latent-logic-using-imitation-learning-to-help-self-driving-cars-learn-human-behavior","status":"publish","type":"post","link":"https:\/\/www.rudebaguette.com\/en\/2019\/12\/waymo-acquires-the-uk-firm-latent-logic-using-imitation-learning-to-help-self-driving-cars-learn-human-behavior\/","title":{"rendered":"Waymo acquires the UK firm Latent Logic, using ‘imitation learning’ to help self-driving cars learn human behavior"},"content":{"rendered":"\n

Waymo has acquired the UK artificial intelligence firm Latent Logic, giving Alphabet\u2019s self-driving car company its first foothold in the UK, according to The Guardian<\/em><\/a>. <\/p>\n\n\n\n

With roots at Oxford University, Latent Logic specializes in a type of machine learning called imitation learning, which models the behavior of human drivers and pedestrians to help emulate and react to their actions on the road. Waymo says the approach will help its AI learn how to navigate complex real-world situations. <\/p>\n\n\n\n

Many machine learning platforms rely on reinforcement learning, in which an engineer flags desired behavior to be repeated in future situations, according to Forbes<\/em><\/a>. But problems can occur when AI encounters unpredictable situations in the real world that may not fit with these learning scenarios. Reinforcement learning can also encourage reward-maximizing behavior instead of the most practical and realistic response to complex, real-world problems.  <\/p>\n\n\n\n

Founded in 2017, Latent Logic uses footage from traffic cameras to model the behavior of human drivers, pedestrians, and cyclists. Waymo will now use this modeling to leverage data collected from 10 million miles of driving from its 600 self-driving vehicles. <\/p>\n\n\n\n

\u201cThe team\u2019s expertise in reinforcement learning and imitation learning can help further accelerate Waymo\u2019s progress in areas from simulation to behaviour prediction and planning,\u201d Waymo said in a statement. <\/p>\n\n\n\n

Latent Logic founders Shimon Whiteson and Jo\u00e3o Messia will join Waymo, and the Latent Logic team will stay in Oxford. Waymo hopes to develop an additional engineering hub and talent base in the UK, home to some of the world\u2019s most advanced AI research. The move will allow Waymo to hire researchers that wouldn\u2019t have relocated to their headquarters in California. <\/p>\n\n\n\n

Waymo hasn\u2019t disclosed the amount paid to acquire Latent Logic.\u00a0The company says it\u2019s just the beginning of their plans to invest in Europe and the UK. Earlier this year, they launched\u00a0partnerships with Renault and Nissan<\/a>\u00a0to research the viability of autonomous vehicles in France and Japan. In October, they announced plans with Renault to investigate options to establish an autonomous transportation route in Paris.\u00a0<\/p>\n\n\n\n

Whiteson, chief scientist at Latent Logic, said:<\/p>\n\n\n\n

\u201cBy joining Waymo, we are taking a big leap towards realising our ambition of safe, self-driving vehicles. In just two years, we have made significant progress in using imitation learning to simulate real human behaviours on the road. I\u2019m excited by what we can now achieve in combining this expertise with the talent, resources and progress Waymo have already made in self-driving technology.\u201d<\/p>\n\n\n\n

Photo by Grendelkhan [CC BY-SA 4.0 (https:\/\/creativecommons.org\/licenses\/by-sa\/4.0)]<\/p>\n","protected":false},"excerpt":{"rendered":"

Waymo has acquired the UK artificial intelligence firm Latent Logic, giving Alphabet\u2019s self-driving car company its first foothold in the UK, according to The Guardian.  With roots at Oxford University, Latent Logic specializes in a type of machine learning called imitation learning, which models the behavior of human drivers and pedestrians to help emulate and react to<\/p>\n","protected":false},"author":77,"featured_media":48091,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"subtitle":"","footnotes":""},"categories":[113],"tags":[1522],"class_list":["post-48090","post","type-post","status-publish","format-standard","has-post-thumbnail","category-finance","tag-artificial-intelligence"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/posts\/48090","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/users\/77"}],"replies":[{"embeddable":true,"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/comments?post=48090"}],"version-history":[{"count":0,"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/posts\/48090\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/media\/48091"}],"wp:attachment":[{"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/media?parent=48090"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/categories?post=48090"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.rudebaguette.com\/en\/wp-json\/wp\/v2\/tags?post=48090"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}