{"id":24049,"date":"2016-03-28T17:56:39","date_gmt":"2016-03-29T00:56:39","guid":{"rendered":"http:\/\/lifeboat.com\/blog\/2016\/03\/ibm-wants-to-accelerate-ai-learning-with-new-processor-tech"},"modified":"2017-06-04T20:04:28","modified_gmt":"2017-06-05T03:04:28","slug":"ibm-wants-to-accelerate-ai-learning-with-new-processor-tech","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2016\/03\/ibm-wants-to-accelerate-ai-learning-with-new-processor-tech","title":{"rendered":"IBM wants to accelerate AI learning with new processor tech"},"content":{"rendered":"<p><a class=\"blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/ibm-wants-to-accelerate-ai-learning-with-new-processor-tech.jpg\"><\/a><\/p>\n<p>Deep neural networks (DNNs) can be taught nearly anything, including how to <a href=\"http:\/\/www.engadget.com\/2016\/03\/14\/the-final-lee-sedol-vs-alphago-match-is-about-to-start\/\">beat us<\/a> at our own games. The problem is that training <a href=\"http:\/\/www.engadget.com\/2016\/03\/14\/microsoft-wants-you-to-train-ai-with-minecraft\/\">AI systems<\/a> ties up big-ticket supercomputers or data centers for days at a time. Scientists from IBM\u2019s T.J. Watson Research Center think they can cut the horsepower and learning times drastically using \u201c<a href=\"http:\/\/arxiv.org\/ftp\/arxiv\/papers\/1603\/1603.07341.pdf\">resistive processing units<\/a>,\u201d theoretical chips that combine CPU and non-volatile memory. Those could accelerate data speeds exponentially, resulting in systems that can do tasks like \u201cnatural speech recognition and translation between all world languages,\u201d according to the team.<\/p>\n<p>So why does it take so much computing power and time to teach AI? The problem is that modern neural networks like Google\u2019s <a href=\"http:\/\/www.engadget.com\/2016\/02\/05\/google-deepmind-ai-finds-its-way-through-a-3D-maze-by-sight\/\">DeepMind<\/a> or <a href=\"http:\/\/www.engadget.com\/2016\/02\/22\/ibm-emotion-detection-upgrade\/\">IBM Watson<\/a> must perform billions of tasks in in parallel. That requires numerous CPU memory calls, which quickly adds up over billions of cycles. The researchers debated using new storage tech like <a href=\"http:\/\/www.engadget.com\/2014\/07\/24\/future-phone-terabyte-storage\/\">resistive RAM<\/a> that can permanently store data with DRAM-like speeds. However, they eventually came up with the idea for a new type of chip called a resistive processing unit (RPU) that puts large amounts of resistive RAM directly onto a CPU.<\/p>\n<p><!-- Link: <a href=\"http:\/\/www.engadget.com\/2016\/03\/28\/ibm-resistive-processing-deep-learning\/\">http:\/\/www.engadget.com\/2016\/03\/28\/ibm-resistive-processing-deep-learning\/<\/a> --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Deep neural networks (DNNs) can be taught nearly anything, including how to beat us at our own games. The problem is that training AI systems ties up big-ticket supercomputers or data centers for days at a time. Scientists from IBM\u2019s T.J. Watson Research Center think they can cut the horsepower and learning times drastically using [\u2026]<\/p>\n","protected":false},"author":367,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6,44],"tags":[],"class_list":["post-24049","post","type-post","status-publish","format-standard","hentry","category-robotics-ai","category-supercomputing"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/24049","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\/367"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=24049"}],"version-history":[{"count":3,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/24049\/revisions"}],"predecessor-version":[{"id":67887,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/24049\/revisions\/67887"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=24049"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=24049"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=24049"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}