{"id":178920,"date":"2024-08-12T10:46:52","date_gmt":"2024-08-12T14:46:52","guid":{"rendered":"https:\/\/inspenet.com\/noticias\/google-deepmind-and-the-table-tennis-robot\/"},"modified":"2024-08-12T20:28:33","modified_gmt":"2024-08-13T00:28:33","slug":"google-deepmind-and-the-table-tennis-robot","status":"publish","type":"noticias","link":"https:\/\/inspenet.com\/en\/news\/google-deepmind-and-the-table-tennis-robot\/","title":{"rendered":"Google DeepMind develops table tennis robot that manages to &#8220;beat&#8221; humans"},"content":{"rendered":"\n<p>Over the past four decades, humans have maintained an advantage over the table tennis robot.\nHowever, recent studies by Google DeepMind may be changing the landscape.\nAccording to a preliminary paper published Aug. 7, researchers have developed a robotic system that has achieved a level of performance comparable to that of an &#8220;amateur ping pong player.&#8221;  <\/p>\n\n<h2 class=\"wp-block-heading\">Google DeepMind studies<\/h2>\n\n<p>Traditionally, games such as chess and Go have been the games of choice for evaluating the strategic capabilities of artificial intelligence.\nHowever, table tennis has become a standard in the robotics industry due to its unique combination of real-time strategy and physicality.\nThe sport requires rapid adaptation to dynamic variables, complex movements and visual coordination, which has led engineers to engage in countless rounds of ping pong with machines for more than 40 years.  <\/p>\n\n<blockquote class=\"twitter-tweet\"><p lang=\"en\" dir=\"ltr\">Meet our AI-powered robot that&#8217;s ready to play table tennis.\n\ud83e\udd16\ud83c\udfd3   <br\/><br\/>It&#8217;s the first agent to achieve amateur human level performance in this sport.\nHere&#8217;s how it works.\n\ud83e\uddf5 <a href=\"https:\/\/t.co\/AxwbRQwYiB\">pic.twitter.com\/AxwbRQwYiB<\/a>  <\/p>&#8211; Google DeepMind (@GoogleDeepMind) <a href=\"https:\/\/twitter.com\/GoogleDeepMind\/status\/1821562365931855970?ref_src=twsrc%5Etfw\" target=\"_blank\" rel=\"noopener\">August 8, 2024<\/a><\/blockquote> <script async=\"\" src=\"https:\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script>\n\n<p>In an announcement on the X platform, <strong>Google DeepMind explained that the robot had to master basic skills<\/strong>, such as ball reception, as well as more advanced skills, including strategizing and long-term planning to achieve a goal.\nTo achieve this, engineers began by collecting an extensive dataset of &#8220;initial states of table tennis balls,&#8221; which contained detailed information about the position, spin and velocity of the balls. <\/p>\n\n<p>With this data, the artificial intelligence practiced in highly accurate virtual simulations, learning techniques such as returning serves, aiming the backhand and applying<em>topspin <\/em>on the forehand.\n <strong>They<\/strong> then <strong>integrated the <a href=\"https:\/\/inspenet.com\/en\/news\/robotic-arm-to-clean-substacons\/\" target=\"_blank\" data-type=\"post\" data-id=\"146073\" rel=\"noreferrer noopener\">AI with a robotic arm<\/a><\/strong>, capable of executing complex and fast movements, and pitted it against human players. <\/p>\n\n<h2 class=\"wp-block-heading\">Machine learning of the table tennis robot<\/h2>\n\n<p>This training process did not end there.\nThe visual information captured by the robot&#8217;s cameras was analyzed again in simulations,<strong> creating a continuous cycle of feedback and<\/strong> <strong>machine<\/strong><strong> learning<\/strong>.\nThis cycle allowed the system to adapt and constantly improve its performance.  <\/p>\n\n<p>Finally, it was time to put the robot to the test in a tournament.\nThe project&#8217;s researchers called 29 human players, classified into four skill levels: <em>beginner, intermediate, advanced and &#8220;advanced+&#8221;.<\/em> The robotic arm faced each of them, winning 13 of the matches, representing 45% of the time.\nAccording to the researchers, these results demonstrate that the robot has achieved solid performance at the amateur level.  <\/p>\n<div class=\" align\"><div id=\"wpfp_2959\" data-item=\"{&quot;sources&quot;:[{&quot;src&quot;:&quot;https:\\\/\\\/www.youtube.com\\\/watch?v=EqQl-JQxToE&quot;,&quot;type&quot;:&quot;video\\\/youtube&quot;,&quot;sources_fvqs&quot;:[{&quot;src&quot;:&quot;https:\\\/\\\/www.youtube.com\\\/watch?v=EqQl-JQxToE&quot;,&quot;type&quot;:&quot;video\\\/youtube&quot;},{&quot;src&quot;:&quot;https:\\\/\\\/www.youtube.com\\\/watch&quot;,&quot;type&quot;:&quot;video\\\/youtube&quot;},{&quot;src&quot;:&quot;https:\\\/\\\/www.youtube.com\\\/watch&quot;,&quot;type&quot;:&quot;video\\\/youtube&quot;},{&quot;src&quot;:&quot;https:\\\/\\\/www.youtube.com\\\/watch&quot;,&quot;type&quot;:&quot;video\\\/youtube&quot;},{&quot;src&quot;:&quot;https:\\\/\\\/www.youtube.com\\\/watch&quot;,&quot;type&quot;:&quot;video\\\/youtube&quot;}]}],&quot;id&quot;:&quot;3738&quot;,&quot;splash&quot;:&quot;https:\\\/\\\/i.ytimg.com\\\/vi\\\/EqQl-JQxToE\\\/maxresdefault.jpg#auto&quot;}\" class=\"freedomplayer flowplayer no-brand is-splash is-paused skin-custom no-svg fp-full fp-outlined is-youtube is-youtube-standard\" data-fvsticky=\"true\" data-button-repeat=\"1\" data-button-rewind=\"1\" data-fv-embed=\"https:\/\/inspenet.com\/en\/news\/google-deepmind-and-the-table-tennis-robot\/fvp-2959\/\" style=\"position:relative; max-width: 95%; max-height: 90%; \" data-ratio=\"0.5625\" data-qsel=\"1080p,720p,480p,360p\" data-qlabels=\"1080p,720p,480p,360p\" data-fv_stats_data=\"{&quot;player_id&quot;:&quot;2959&quot;,&quot;post_id&quot;:178920}\" data-engine=\"fvyoutube\" data-speedb=\"1\">\n\t<div class=\"fp-ratio\" style=\"padding-top: 56.25%\"><\/div>\n\t<img decoding=\"async\" style=\"position:absolute;top:0;left:0;width:100%\" class=\"fp-splash\" alt=\"video\" src=\"https:\/\/i.ytimg.com\/vi\/EqQl-JQxToE\/maxresdefault.jpg#auto\" \/>\n\t<div class=\"fp-ui\" style=\"position:absolute\"><noscript>Please enable JavaScript<\/noscript><div class=\"fp-play fp-visible\"><svg class=\"fp-play-rounded-outline\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" viewBox=\"0 0 99.844 99.8434\"><defs><style>.fp-color-play{opacity:0.65;}.controlbutton{fill:#fff;}<\/style><\/defs><title>play-rounded-outline<\/title><path class=\"fp-color-play\" d=\"M49.9217-.078a50,50,0,1,0,50,50A50.0564,50.0564,0,0,0,49.9217-.078Z\"\/><path class=\"controlbutton\" d=\"M41.0359,71.19a5.0492,5.0492,0,0,1-2.5575-.6673c-1.8031-1.041-2.7958-3.1248-2.7958-5.8664V35.1887c0-2.7429.9933-4.8272,2.797-5.8676,1.8025-1.0422,4.1034-.86,6.48.5143L70.4782,44.5672c2.3751,1.3711,3.6826,3.2725,3.6832,5.3545s-1.3076,3.9845-3.6832,5.3562L44.9592,70.0114A7.9384,7.9384,0,0,1,41.0359,71.19Zm.0065-40.123a2.6794,2.6794,0,0,0-1.3582.3413c-1.0263.5926-1.5912,1.9349-1.5912,3.78V64.6563c0,1.8449.5649,3.1866,1.5906,3.7791,1.0281.5932,2.4733.4108,4.07-.512L69.273,53.1906c1.5983-.9227,2.478-2.0838,2.478-3.2689s-.88-2.3445-2.478-3.2666L43.754,31.9227A5.5685,5.5685,0,0,0,41.0423,31.0671Z\" filter=\"url(#f1)\"\/><\/svg><\/div><div class=\"fp-preload\"><b><\/b><b><\/b><b><\/b><b><\/b><\/div><\/div>\n<div class='fvp-share-bar'><div><label><a class=\"embed-code-toggle\" href=\"#\"><strong>Embed<\/strong><\/a><\/label><\/div><div class=\"embed-code\"><label>Copy and paste this HTML code into your webpage to embed.<\/label><textarea><\/textarea><\/div><\/div>\n<\/div>\n<\/div>\n<p><\/p>\n\n<p class=\"has-text-align-center\">  Some highlights, among the human level of competition in table tennis with robots.\nSource: Atil Iscen <\/p>\n\n<p>Those passionate about table tennis who fear losing their superiority to robots <strong>can rest easy<\/strong><em>(at least for now<\/em>).\nAlthough the machine beat all beginner-level players, its win rate dropped to 55% against intermediate players, and it failed to win against advanced players.\nParticipants in the study described the experience as &#8220;fun&#8221; and &#8220;entertaining,&#8221; regardless of the outcome, and showed a strong interest in having rematches with the robot.  <\/p>\n\n<p><strong>Follow us on social networks and don&#8217;t miss any of our publications!<\/strong><\/p>\n\n<p><a href=\"https:\/\/www.youtube.com\/@inspenet\" target=\"_blank\" rel=\"noreferrer noopener\">YouTube<\/a> <a href=\"https:\/\/www.linkedin.com\/company\/inspenetnetwork\" target=\"_blank\" rel=\"noreferrer noopener\">LinkedIn<\/a> <a href=\"https:\/\/www.facebook.com\/inspenetnetwork\" target=\"_blank\" rel=\"noreferrer noopener\">Facebook<\/a> <a href=\"https:\/\/www.instagram.com\/inspenet\/\" target=\"_blank\" rel=\"noreferrer noopener\">Instagram<\/a> <a href=\"https:\/\/twitter.com\/Inspenetnetwork\" target=\"_blank\" rel=\"noreferrer noopener\">X<\/a><\/p>\n\n<p><strong>Source and photo:<\/strong> <strong><a href=\"https:\/\/sites.google.com\/view\/competitive-robot-table-tennis\/home\" target=\"_blank\" rel=\"noreferrer noopener\">Google DeepMind<\/a><\/strong><\/p>\n","protected":false},"excerpt":{"rendered":"<p>  Google&#8217;s DeepMind robot won 45% of its matches against human players, but failed to beat advanced level players.<\/p>\n","protected":false},"author":8404,"featured_media":178751,"menu_order":0,"comment_status":"open","ping_status":"open","template":"","meta":{"_acf_changed":false,"_bbp_topic_count":0,"_bbp_reply_count":0,"_bbp_total_topic_count":0,"_bbp_total_reply_count":0,"_bbp_voice_count":0,"_bbp_anonymous_reply_count":0,"_bbp_topic_count_hidden":0,"_bbp_reply_count_hidden":0,"_bbp_forum_subforum_count":0,"footnotes":""},"categoria_noticias":[13030,13058],"etiqueta_noticias":[],"class_list":["post-178920","noticias","type-noticias","status-publish","has-post-thumbnail","hentry","categoria_noticias-industry","categoria_noticias-robotics"],"acf":[],"_links":{"self":[{"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/noticias\/178920","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/noticias"}],"about":[{"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/types\/noticias"}],"author":[{"embeddable":true,"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/users\/8404"}],"replies":[{"embeddable":true,"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/comments?post=178920"}],"version-history":[{"count":0,"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/noticias\/178920\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/media\/178751"}],"wp:attachment":[{"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/media?parent=178920"}],"wp:term":[{"taxonomy":"categoria_noticias","embeddable":true,"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/categoria_noticias?post=178920"},{"taxonomy":"etiqueta_noticias","embeddable":true,"href":"https:\/\/inspenet.com\/en\/wp-json\/wp\/v2\/etiqueta_noticias?post=178920"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}