SoylentNews is people
SoylentNews
New quick-learning neural network powered by memristors
A new type of neural network made with memristors can dramatically improve the efficiency of teaching machines to think like humans. The network, called a reservoir computing system, could predict words before they are said during conversation, and help predict future outcomes based on the present.
The research team that created the reservoir computing system, led by Wei Lu, U-M professor of electrical engineering and computer science, recently published their work [open, DOI: 10.1038/s41467-017-02337-y] [DX] in Nature Communications.
Reservoir computing systems, which improve on a typical neural network's capacity and reduce the required training time, have been created in the past with larger optical components. However, the U-M group created their system using memristors, which require less space and can be integrated more easily into existing silicon-based electronics. [...] When a set of data is inputted into the reservoir, the reservoir identifies important time-related features of the data, and hands it off in a simpler format to a second network. This second network then only needs training like simpler neural networks, changing weights of the features and outputs that the first network passed on until it achieves an acceptable level of error. "The beauty of reservoir computing is that while we design it, we don't have to train it," says Lu.
[...] Using only 88 memristors as nodes to identify handwritten versions of numerals, compared to a conventional network that would require thousands of nodes for the task, the reservoir achieved 91% accuracy.
(Score: 1, Interesting) by Anonymous Coward on Sunday December 31 2017, @11:58PM (2 children)
Is seems like there are two separate things here:
1. Reservoir Computing: https://en.wikipedia.org/wiki/Reservoir_computing [wikipedia.org]
2. Memristors implementations of neural networks (and maybe their applicability specifically to Reservoir Computing?)
It seems like the claim is that Reservoir Computing is a design for neural networks that only require training a small part of the network (The "readout" stage), and this was efficient for some standard tests of neural networks.
It is not clear to me if the claim regarding memristors is that they allow you to have less nodes than a neural network somehow (that seems non-nonsensical: alternative hardware does not reduce logical node count), or something else.
(Score: 4, Informative) by takyon on Monday January 01 2018, @12:05AM (1 child)
It's not the first time this has come up:
https://www.nextplatform.com/2015/07/08/memristors-mimic-brains-for-massive-machine-learning/ [nextplatform.com]
https://www.nextbigfuture.com/2015/09/advanced-memristors-are-pathway-to.html [nextbigfuture.com]
https://www.nextbigfuture.com/2016/08/ibm-scientists-imitate-functionality-of_18.html [nextbigfuture.com]
https://www.nextbigfuture.com/2016/02/polymer-memristor-used-as-basis-for.html [nextbigfuture.com]
https://www.nextbigfuture.com/2015/05/proof-of-concept-all-memristor-brain.html [nextbigfuture.com]
And here's something from the same University of Michigan researcher:
https://www.nextbigfuture.com/2013/08/memristor-based-self-organizing.html [nextbigfuture.com]
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: -1, Spam) by Anonymous Coward on Monday January 01 2018, @12:17AM
A man walked up to a woman who was sitting on a park bench, watching her 9-year-old girl play. The man's appearance could only be described as 'creepy' and 'unsettling'; he was cockeyed, wore small glasses, and had a thin mustache. After staring at the woman for a few moments, the man finally spoke these words: "You know, I've been looking it, and... hehe!"
Suddenly, the man instantly changed locations and was sitting on the grass, naked. The mother's child was sitting on top of him, also naked, and with the man's penis entering her. However, something was different about the girl's appearance; she was now incredibly obese, unlike previously; she had countless crater-like indentations on her buttocks, and the skin on the inside of said craters was a sickly greenish color; and finally, the craters on her skin were slowly leaking pus and blood.
The man began thrusting, and every time he did so, he screamed, "Looking it!" These words were shouted again and again, and echoed throughout the park. Occasionally, the man would scream other phrases, such as 'Delicious!', at the same time as he was screaming other words; what was most odd was that it sounded as if he was screaming two entirely different things simultaneously, almost as if someone who sounded exactly like he did was speaking over him. The man's thrusting became increasingly rapid over time, and so too did his shouts. Finally, the man squeezed the girl's buttocks with as much strength as he could muster, and a vomit-inducing mixture of pus and blood spewed everywhere, accompanied by a high-pitched scream
Elsewhere, a naked woman - the child's mother - could be seen hanging from her rectum on the end of an abnormally-large telephone pole which reached thousands of feet tall. The woman's rectum appeared to be stretching out to unbelievable lengths, to the point where it resembled a long hood. From this location, the man's everlasting shouts of pleasure could be heard...