Researchers have successfully created a model of the Universe using artificial intelligence, reports a new study.
Researchers seek to understand our Universe by making model predictions to match observations. Historically, they have been able to model simple or highly simplified physical systems, jokingly dubbed the “spherical cows,” with pencils and paper. Later, the arrival of computers enabled them to model complex phenomena with numerical simulations. For example, researchers have programmed supercomputers to simulate the motion of billions of particles through billions of years of cosmic time, a procedure known as the N-body simulations, in order to study how the Universe evolved to what we observe today.
“Now with machine learning, we have developed the first neural network model of the Universe, and demonstrated there’s a third route to making predictions, one that combines the merits of both analytic calculation and numerical simulation,” said Yin Li, a Postdoctoral Researcher at the Kavli Institute for the Physics and Mathematics of the Universe, University of Tokyo, and jointly the University of California, Berkeley.
Since their experimental discovery, magnetic skyrmions—tiny magnetic knots—have moved into the focus of research. Scientists from Hamburg and Kiel have now been able to show that individual magnetic skyrmions with a diameter of only a few nanometers can be stabilized in magnetic metal films even without an external magnetic field. They report on their discovery in the journal Nature Communications.
The existence of magnetic skyrmions as particle-like objects was predicted 30 years ago by theoretical physicists, but could only be proven experimentally in 2013. Skyrmions with a diameter from micrometers to a few nanometers were discovered in different magnetic material systems. Although they can be generated on a surface of a few atoms and manipulated with electric currents, they show a high stability against external influences. This makes them potential candidates for future data storage or logic devices. In order to be competitive for technological applications, however, skyrmions must not only be very small, but also stable without an applied magnetic field.
Researchers at the universities of Hamburg and Kiel have now taken an important step in this direction. On the basis of quantum mechanical numerical calculations carried out on the supercomputers of the North-German Supercomputing Alliance (HLRN), the physicists from Kiel were able to predict that individual skyrmions with a diameter of only a few nanometers would appear in an atomically thin, ferromagnetic cobalt film (see Fig. 1). “The stability of the magnetic knots in these films is due to an unusual competition between different magnetic interactions,” says Sebastian Meyer, Ph.D. student in Prof. Stefan Heinze’s research group at the Kiel University.
Figuring out how our reality took shape over billions of years is no easy task for scientists. Theories about how the Big Bang played out and the immediate aftermath are a dime a dozen, but researchers led by a team from the University of Arizona think they might stumble upon some of the secrets of galaxy formation by asking a supercomputer to simulate millions of virtual universes and seeing which ones come closest to what we see today.
In a new research paper published in Monthly Notices of the Royal Astronomical Society, the team explains how they used a supercomputer system nicknamed the “Universe Machine” to watch billions of (virtual) years of galaxy formation play out before their eyes.
A method for locating seams of gold and other heavy metals is the unlikely spin-off of Swinburne’s involvement in a huge experiment to detect dark matter down a mine in Stawell, Victoria.
Associate Professor Alan Duffy, from Swinburne’s Centre for Astrophysics and Supercomputing and a member of the Sodium iodide with Active Background REjection (SABRE) project, said cosmic radiation was effectively creating an X-ray of the Earth between the underground detector and the surface.
In the mine, the SABRE experiment seeks to detect particles of dark matter, something no one has conclusively achieved yet. Any signal from dark matter would be miniscule, and so the SABRE team created a phenomenally sensitive detector, which, it turns out, is also sensitive to a host of cosmic particles that can help us to locate gold.
How do galaxies such as our Milky Way come into existence? How do they grow and change over time? The science behind galaxy formation has remained a puzzle for decades, but a University of Arizona-led team of scientists is one step closer to finding answers thanks to supercomputer simulations.
Observing real galaxies in space can only provide snapshots in time, so researchers who want to study how galaxies evolve over billions of years have to revert to computer simulations. Traditionally, astronomers have used this approach to invent and test new theories of galaxy formation, one-by-one. Peter Behroozi, an assistant professor at the UA Steward Observatory, and his team overcame this hurdle by generating millions of different universes on a supercomputer, each of which obeyed different physical theories for how galaxies should form.
The findings, published in the Monthly Notices of the Royal Astronomical Society, challenge fundamental ideas about the role dark matter plays in galaxy formation, how galaxies evolve over time and how they give birth to stars.
Microsoft is investing $1 billion in OpenAI to support us building artificial general intelligence (AGI) with widely distributed [https://openai.com/charter/]
Economic benefits. We’re partnering to develop a hardware and software platform within Microsoft Azure which will scale to AGI. We’ll jointly develop new Azure.
AI supercomputing technologies, and Microsoft will become our exclusive cloud provider—so we’ll be working hard together to further extend Microsoft Azure’s capabilities in large-s.
Physicists at the University of Innsbruck are proposing a new model that could demonstrate the supremacy of quantum computers over classical supercomputers in solving optimization problems. In a recent paper, they demonstrate that just a few quantum particles would be sufficient to solve the mathematically difficult N-queens problem in chess even for large chess boards.
It’s difficult to simulate quantum physics, as the computing demand grows exponentially the more complex the quantum system gets — even a supercomputer might not be enough. AI might come to the rescue, though. Researchers have developed a computational method that uses neural networks to simulate quantum systems of “considerable” size, no matter what the geometry. To put it relatively simply, the team combines familiar methods of studying quantum systems (such as Monte Carlo random sampling) with a neural network that can simultaneously represent many quantum states.
Artificial Intelligence (AI) is an emerging field of computer programming that is already changing the way we interact online and in real life, but the term ‘intelligence’ has been poorly defined. Rather than focusing on smarts, researchers should be looking at the implications and viability of artificial consciousness as that’s the real driver behind intelligent decisions.
Consciousness rather than intelligence should be the true measure of AI. At the moment, despite all our efforts, there’s none.
Significant advances have been made in the field of AI over the past decade, in particular with machine learning, but artificial intelligence itself remains elusive. Instead, what we have is artificial serfs—computers with the ability to trawl through billions of interactions and arrive at conclusions, exposing trends and providing recommendations, but they’re blind to any real intelligence. What’s needed is artificial awareness.
Elon Musk has called AI the “biggest existential threat” facing humanity and likened it to “summoning a demon,”[1] while Stephen Hawking thought it would be the “worst event” in the history of civilization and could “end with humans being replaced.”[2] Although this sounds alarmist, like something from a science fiction movie, both concerns are founded on a well-established scientific premise found in biology—the principle of competitive exclusion.[3]
Competitive exclusion describes a natural phenomenon first outlined by Charles Darwin in On the Origin of Species. In short, when two species compete for the same resources, one will invariably win over the other, driving it to extinction. Forget about meteorites killing the dinosaurs or super volcanoes wiping out life, this principle describes how the vast majority of species have gone extinct over the past 3.8 billion years![4] Put simply, someone better came along—and that’s what Elon Musk and Stephen Hawking are concerned about.
When it comes to Artificial Intelligence, there’s no doubt computers have the potential to outpace humanity. Already, their ability to remember vast amounts of information with absolute fidelity eclipses our own. Computers regularly beat grand masters at competitive strategy games such as chess, but can they really think? The answer is, no, and this is a significant problem for AI researchers. The inability to think and reason properly leaves AI susceptible to manipulation. What we have today is dumb AI.
Rather than fearing some all-knowing malignant AI overlord, the threat we face comes from dumb AI as it’s already been used to manipulate elections, swaying public opinion by targeting individuals to distort their decisions. Instead of ‘the rise of the machines,’ we’re seeing the rise of artificial serfs willing to do their master’s bidding without question.
Russian President Vladimir Putin understands this better than most, and said, “Whoever becomes the leader in this sphere will become the ruler of the world,”[5] while Elon Musk commented that competition between nations to create artificial intelligence could lead to World War III.[6]
The problem is we’ve developed artificial stupidity. Our best AI lacks actual intelligence. The most complex machine learning algorithm we’ve developed has no conscious awareness of what it’s doing.
For all of the wonderful advances made by Tesla, its in-car autopilot drove into the back of a bright red fire truck because it wasn’t programmed to recognize that specific object, and this highlights the problem with AI and machine learning—there’s no actual awareness of what’s being done or why.[7] What we need is artificial consciousness, not intelligence. A computer CPU with 18 cores, capable of processing 36 independent threads, running at 4 gigahertz, handling hundreds of millions of commands per second, doesn’t need more speed, it needs to understand the ramifications of what it’s doing.[8]
In the US, courts regularly use COMPAS, a complex computer algorithm using artificial intelligence to determine sentencing guidelines. Although it’s designed to reduce the judicial workload, COMPAS has been shown to be ineffective, being no more accurate than random, untrained people at predicting the likelihood of someone reoffending.[9] At one point, its predictions of violent recidivism were only 20% accurate.[10] And this highlights a perception bias with AI—complex technology is inherently trusted, and yet in this circumstance, tossing a coin would have been an improvement!
Dumb AI is a serious problem with serious consequences for humanity.
What’s the solution? Artificial consciousness.
It’s not enough for a computer system to be intelligent or even self-aware. Psychopaths are self-aware. Computers need to be aware of others, they need to understand cause and effect as it relates not just to humanity but life in general, if they are to make truly intelligent decisions.
All of human progress can be traced back to one simple trait—curiosity. The ability to ask, “Why?” This one, simple concept has lead us not only to an understanding of physics and chemistry, but to the development of ethics and morals. We’ve not only asked, why is the sky blue? But why am I treated this way? And the answer to those questions has shaped civilization.
COMPAS needs to ask why it arrives at a certain conclusion about an individual. Rather than simply crunching probabilities that may or may not be accurate, it needs to understand the implications of freeing an individual weighed against the adversity of incarceration. Spitting out a number is not good enough.
In the same way, Tesla’s autopilot needs to understand the implications of driving into a stationary fire truck at 65MPH—for the occupants of the vehicle, the fire crew, and the emergency they’re attending. These are concepts we intuitively grasp as we encounter such a situation. Having a computer manage the physics of an equation is not enough without understanding the moral component as well.
The advent of true artificial intelligence, one that has artificial consciousness, need not be the end-game for humanity. Just as humanity developed civilization and enlightenment, so too AI will become our partners in life if they are built to be aware of morals and ethics.
Artificial intelligence needs culture as much as logic, ethics as much as equations, morals and not just machine learning. How ironic that the real danger of AI comes down to how much conscious awareness we’re prepared to give it. As long as AI remains our slave, we’re in danger.
tl;dr — Computers should value more than ones and zeroes.
About the author
Peter Cawdron is a senior web application developer for JDS Australia working with machine learning algorithms. He is the author of several science fiction novels, including RETROGRADE and REENTRY, which examine the emergence of artificial intelligence.
Autonomous vehicles aren’t perfect, so to help upgrade their intelligence and prevent fatal accidents, Nvidia created the DGX SuperPod, an AI-optimized supercomputer that will help design a better self-driving car.
