{"id":25045,"date":"2014-08-08T14:00:52","date_gmt":"2014-08-08T12:00:52","guid":{"rendered":"https:\/\/ocasapiens-dweb.blogautore.repubblica.it\/?p=25045"},"modified":"2021-10-10T14:21:05","modified_gmt":"2021-10-10T12:21:05","slug":"neurofuture","status":"publish","type":"post","link":"https:\/\/archivio.ocasapiens.org\/index.php\/2014\/08\/08\/neurofuture\/","title":{"rendered":"Neurofuture?"},"content":{"rendered":"<p>Il punto interrogativo l&#8217;ho aggiunto io. \u00a0<strong><a href=\"http:\/\/neurofuture.eu\/\" target=\"_blank\" rel=\"noopener\">Neurofuture<\/a><\/strong> &#8211; ricordo ai distratti &#8211; \u00e8 il sito degli scienziati che contestano il programma per lo Human Brain Project presentato a giugno da Henry Markram e che la direzione per l&#8217;Innovazione tecnologica della Commissione Europea ha promesso di modificare prima di approvarlo a dicembre.<br \/>\nGuarda caso, su <em>Science<\/em> di oggi un paper descrive un&#8217;innovazione che figura tra quelle potenzialmente derivabili dai risultati &#8211; sempre che nel 2023 ci siano &#8211; di quel &#8220;progetto bandiera&#8221; da 1,2 miliardi di euro.<\/p>\n<p>Per realizzarla, i <strong><a href=\"http:\/\/www.sciencemag.org\/content\/345\/6197\/668.abstract\" target=\"_blank\" rel=\"noopener\">ricercatori <\/a><\/strong>dell&#8217;IBM e di <strong><a href=\"http:\/\/tech.cornell.edu\/\" target=\"_blank\" rel=\"noopener\">Cornell Tech<\/a>\u00a0<\/strong>non hanno avuto bisogno di simulare l&#8217;intero cervello umano in un super-computer:<\/p>\n<p><em>\u00a0Inspired by the brain\u2019s structure, we have developed an efficient, scalable, and flexible non\u2013von Neumann architecture that leverages contemporary silicon technology. To demonstrate, we built a 5.4-billion-transistor chip with 4096 neurosynaptic cores interconnected via an intrachip network that integrates 1 million programmable spiking neurons and 256 million configurable synapses. Chips can be tiled in two dimensions via an interchip communication interface, seamlessly scaling the architecture to a cortex-like sheet of arbitrary size.<\/em><\/p>\n<p>E nell&#8217;ultima frase dell&#8217;abstract, rigirano il coltello nella piaga:<\/p>\n<p><em>The architecture is well suited to many applications that use complex neural networks in real time, for example, multiobject detection and classification. With 400-pixel-by-240-pixel video input at 30 frames per second, the chip consumes 63 milliwatts.<\/em><\/p>\n<p>Nel caso non fosse ancora chiaro, Robert Service aggiunge<\/p>\n<p><em>All computer chips made today rely on the same general architecture that was outlined nearly 70 years ago. This architecture separates the two primary tasks a chip needs to carry out\u2014processing and memory\u2014into different regions and continuously communicates data back and forth. Though this strategy works well for crunching numbers and running spreadsheets, it&#8217;s much less efficient for handling tasks that manage vast amounts of data, such as vision and language processing. But in recent years, researchers around the globe have been pursuing a new approach called <strong><a href=\"http:\/\/www.economist.com\/news\/science-and-technology\/21582495-computers-will-help-people-understand-brains-better-and-understanding-brains\" target=\"_blank\" rel=\"noopener\">neuromorphic computing<\/a><\/strong>. &#8230;.\u00a0The novel chips <strong>could revolutionize<\/strong> efforts in everything from helping computers and robots sense their environment to offering new tools to help blind people navigate their surroundings.<\/em><\/p>\n<p>&#8220;Recent years&#8221; va inteso come dalla met\u00e0 circa degli anni &#8217;80. Aggiunti link e grassett0 al condizionale, ma all&#8217;ETH e all&#8217;universit\u00e0 di Zurigo, il gruppo di <strong><a href=\"http:\/\/www.ini.uzh.ch\/people\/giacomo\" target=\"_blank\" rel=\"noopener\">Giacomo Indiveri<\/a> <\/strong>sta gi\u00e0 facendo chip cos\u00ec per <strong><a href=\"http:\/\/www.ini.uzh.ch\/research\/41661\" target=\"_blank\" rel=\"noopener\">impianti cocleari<\/a><\/strong>.<br \/>\nNon so voi, ma conoscendo i miei neuroni, preferisco un laptop la cui un&#8217;architettura sta in pe&#8217; da 70 anni.<br \/>\n<a href=\"http:\/\/www-03.ibm.com\/press\/us\/en\/pressrelease\/44529.wss\" target=\"_blank\" rel=\"noopener\">Com. stampa<\/a> IBM<\/p>\n<p style=\"text-align: center;\">*<\/p>\n<p><strong><a href=\"http:\/\/www.sciencemag.org\/content\/345\/6197\/644\" target=\"_blank\" rel=\"noopener\">Una collaborazione Harvard-MIT<\/a> <\/strong>ha costruito invece dei robottini con un foglio di carta e un <a href=\"http:\/\/www.samstoybox.com\/toypics\/ShrinkyDinks.jpg\" target=\"_blank\" rel=\"noopener\">giocattolo<\/a> in plastica a memoria di forma. Fanno tutto da soli, si piegano a origami e nel giro di 4 minuti si mettono a strisciare:<br \/>\n<img decoding=\"async\" src=\"http:\/\/www.seas.harvard.edu\/sites\/default\/files\/news\/images\/Image2_op.jpg\" alt=\"\" \/><br \/>\n<a href=\"http:\/\/www.seas.harvard.edu\/news\/2014\/08\/robot-folds-itself-up-and-walks-away\" target=\"_blank\" rel=\"noopener\">Com. stampa<\/a>\u00a0Harvard.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Il punto interrogativo l&#8217;ho aggiunto io. \u00a0Neurofuture &#8211; ricordo ai distratti &#8211; \u00e8 il sito degli scienziati che contestano il programma per lo Human Brain Project presentato a giugno da Henry Markram e che la direzione per l&#8217;Innovazione tecnologica della Commissione Europea ha promesso di modificare prima di approvarlo a dicembre. Guarda caso, su Science&hellip; <a class=\"more-link\" href=\"https:\/\/archivio.ocasapiens.org\/index.php\/2014\/08\/08\/neurofuture\/\">Continua a leggere <span class=\"screen-reader-text\">Neurofuture?<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[19,22,30],"tags":[3280,3299,4781,6014],"class_list":["post-25045","post","type-post","status-publish","format-standard","hentry","category-matematica","category-neuroscienze","category-tecno","tag-hbp","tag-henry-markram","tag-neuromorphic-computing","tag-robotica","entry"],"_links":{"self":[{"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/posts\/25045","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/comments?post=25045"}],"version-history":[{"count":2,"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/posts\/25045\/revisions"}],"predecessor-version":[{"id":42522,"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/posts\/25045\/revisions\/42522"}],"wp:attachment":[{"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/media?parent=25045"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/categories?post=25045"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/archivio.ocasapiens.org\/index.php\/wp-json\/wp\/v2\/tags?post=25045"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}