{"id":1248,"date":"2019-02-04T09:39:01","date_gmt":"2019-02-04T17:39:01","guid":{"rendered":"http:\/\/www.futuretechnology500.com\/?p=1248"},"modified":"2019-02-04T09:39:04","modified_gmt":"2019-02-04T17:39:04","slug":"the-scoop-on-artificial-synapses","status":"publish","type":"post","link":"https:\/\/www.futuretechnology500.com\/index.php\/future-medical-technology\/the-scoop-on-artificial-synapses\/","title":{"rendered":"The scoop on artificial synapses"},"content":{"rendered":"\n

For some time now, we\u2019ve been hearing\npeople tell us that, soon enough, we\u2019ll have circuitry more powerful and more\nefficient than the human mind. It looks like we might finally be on the edge of\nthis becoming true, too: circuitry inspired by biology is in the process of\nbeing created to help create low-powered AI chips. <\/p>\n\n\n\n

\"Artificial<\/figure>\n\n\n\n

While this is still very much in an\nearly developmental phase, the hope is that \u2013 with some hurdles still to\novercome \u2013 we could start creating AI chips with far less demand than we presently\ncan.<\/p>\n\n\n\n

Indeed, this latest finding from the\nUS National Institute of Standards and Technology has created something very\nimpressive indeed. They have developed a new magnetically controlled electronic\nsynapse<\/a>. This new synapse is\ngoing to be made like an AI version of the very ones which link neurons within\nthe mind. However, they are much faster than our brains: millions of times\nfaster, in fact.<\/p>\n\n\n\n

With using around 1\/1000th<\/sup>\nof the energy of the brain, too, this is by far and away the most efficient \u2013\nin both performance and power use \u2013 artificial synapse to date. With the\nability to gather so many signals and fire out electronic pulses all at once,\nthis could be a much more impressive alternative to the more traditional\ntransistor.<\/p>\n\n\n\n

Created by using neuromorphic chips,\nthese will work very much in the similar function to the human brain. This is\nnaturally causing immense excitement within the scientific community. This\nwould allow for us to easily run things like artificial networks \u2013 a key\nconstruct of modern AI development \u2013 with even greater control and efficiency\nthan ever before.<\/p>\n\n\n\n

Best of all? It would allow for vast\nincreases and changes with regards to energy efficiency: the results could be\nabsolutely outstanding.<\/p>\n\n\n\n

Still\nsome work to be done<\/strong><\/h2>\n\n\n\n

However, as you may expect, such\nhardware still has a veritable glut of issues to be worked out and corrected.\nFor sample, they are yet to be upscaled to work within a large device, so it\u2019s\nhard to be clear about whether or not a chip could be built onto one. Also,\nthey are going to be running at temperatures of around 5K at present: no normal\ncomputing practice would allow for such conditions.<\/p>\n\n\n\n

So, while the development of a new\nartificial synapse does feel closer than ever, it\u2019s also pretty clear that it\u2019s\nstill something very much for the decade(s) to come. <\/p>\n","protected":false},"excerpt":{"rendered":"

For some time now, we\u2019ve been hearing people tell us that, soon enough, we\u2019ll have circuitry more powerful and more efficient than the human mind. It looks like we might finally be on the edge of this becoming true, too: circuitry inspired by biology is in the process of being created to help create low-powered […]<\/a><\/p>\n","protected":false},"author":8668,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[18],"tags":[],"class_list":["post-1248","post","type-post","status-publish","format-standard","hentry","category-future-medical-technology","category-18-id","post-seq-1","post-parity-odd","meta-position-corners","fix"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t