SoylentNews is people
SoylentNews
A new theoretical framework is allowing neural networks to learn and recognize patterns on geometric surfaces.
Neural networks based on the visual cortex, called
"convolutional neural networks" (CNNs) have proved surprisingly adept at learning patterns in two-dimensional data—especially in computer vision tasks like recognizing handwritten words and objects in digital images.
CNNs however, are largely stuck in two dimensions.
Now, researchers have delivered, with a new theoretical framework for building neural networks that can learn patterns on any kind of geometric surface. These "gauge-equivariant convolutional neural networks," or gauge CNNs, developed at the University of Amsterdam and Qualcomm AI Research by Taco Cohen, Maurice Weiler, Berkay Kicanaoglu and Max Welling, can detect patterns not only in 2D arrays of pixels, but also on spheres and asymmetrically curved objects. "This framework is a fairly definitive answer to this problem of deep learning on curved surfaces," Welling said.
Applications envisioned include climate modeling, medical scan analysis, computer vision, particle interactions etc.
A gauge CNN would theoretically work on any curved surface of any dimensionality, but Cohen and his co-authors have tested it on global climate data, which necessarily has an underlying 3D spherical structure. They used their gauge-equivariant framework to construct a CNN trained to detect extreme weather patterns, such as tropical cyclones, from climate simulation data. In 2017, government and academic researchers used a standard convolutional network to detect cyclones in the data with 74% accuracy; last year, the gauge CNN detected the cyclones with 97.9% accuracy. (It also outperformed a less general geometric deep learning approach designed in 2018 specifically for spheres — that system was 94% accurate.)
Physicists such as Kyle Cranmer at New York University, are already planning to put Gauge CNNs to work in analyzing four dimensional data related to the forces at work within a proton "a perfect use case for neural networks that have this gauge equivariance."
(Score: 0) by Anonymous Coward on Tuesday January 14 2020, @01:11PM (2 children)
So, Russians developed fission propulsion for/and hypersonic missiles, the Chinese are all over the science articles and let US and EU behind (since 3 years ago, no less) [enago.com], the Nethers now produces AI that can do science (even for newyorkers). Yet some speak about "overall U.S. leadership in science" and "a golden age for innovation in America" [soylentnews.org]. Rrrrright.
If that's how gold looks lately, I'm afraid to think what silver looks like, not to speak about irony.
(Score: 3, Informative) by takyon on Tuesday January 14 2020, @01:26PM (1 child)
That's not confirmed. They may be years away from deployment. Hypersonic won't change anything about mutually assured destruction because you could still detect the missiles or launch a response from submarines. Those missiles will do little to improve the lives of Russians.
They are full of junk science:
Research Fraud at the Highest Levels in China [soylentnews.org]
Low Quality Studies Belie Hype about Research Boom in China [scientificamerican.com]
China’s bevy of supercomputers goes unused [marketwatch.com]
Even if things were better, you would expect a nation with over 4x the population and a comparable or better GDP to produce better results than its rival.
‘The Netherlands is losing ground in artificial intelligence’ [cursor.tue.nl]
Whatever you think the "nethers" are achieving, it will be copied by others. And we haven't even reached the point of neuromorphic "strong AI" doing science research yet.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2, Insightful) by RandomFactor on Tuesday January 14 2020, @09:57PM
While true enough, remember that MAD is a paper tiger.
However an area where hypersonic technology is of dramatic importance is in putting conventional force projection at asymmetrical risk. If a carrier group cannot counter a threat it is at risk in a way the US has little appetite for. If the threat is a missile that can be launched from the air, that's a long ways off indeed. So for a few million dollars you hold off a hundred billion dollar carrier group.
В «Правде» нет известий, в «Известиях» нет правды