Neural take on quantum entanglement

An artist's rendering of a neural network with two layers. At the top is a real quantum system, like atoms in an optical lattice. Below is a network of hidden neurons that capture their interactions. (Credit: E. Edwards/JQI)

Machine learning, the field that’s driving a revolution in artificial intelligence, has cemented its role in modern technology. Its tools and techniques have led to rapid improvements in everything from self-driving cars and speech recognition to the digital mastery of an ancient board game.

Now, physicists are beginning to use machine learning tools to tackle a different kind of problem, one at the heart of quantum physics. In a paper published recently in Physical Review X, researchers from JQI and the Condensed Matter Theory Center (CMTC) at the University of Maryland showed that certain neural networks—abstract webs that pass information from node to node like neurons in the brain—can succinctly describe wide swathes of quantum systems.

Dongling Deng, a JQI Postdoctoral Fellow who is a member of CMTC and the paper’s first author, says that researchers who use computers to study quantum systems might benefit from the simple descriptions that neural networks provide. “If we want to numerically tackle some quantum problem,” Deng says, “we first need to find an efficient representation.”

On paper and, more importantly, on computers, physicists have many ways of representing quantum systems. Typically these representations comprise lists of numbers describing the likelihood that a system will be found in different quantum states. But it becomes difficult to extract properties or predictions from a digital description as the number of quantum particles grows, and the prevailing wisdom has been that entanglement—an exotic quantum connection between particles—plays a key role in thwarting simple representations.

The neural networks used by Deng and his collaborators—CMTC Director and JQI Fellow Sankar Das Sarma and Fudan University physicist and former JQI Postdoctoral Fellow Xiaopeng Li—can efficiently represent quantum systems that harbor lots of entanglement, a surprising improvement over prior methods.

What’s more, the new results go beyond mere representation. “This research is unique in that it does not just provide an efficient representation of highly entangled quantum states,” Das Sarma says. “It is a new way of solving intractable, interacting quantum many-body problems that uses machine learning tools to find exact solutions.”

Neural networks and their accompanying learning techniques powered AlphaGo, the computer program that beat some of the world’s best Go players last year (and the top player this year. The news excited Deng, an avid fan of the board game. Last year, around the same time as AlphaGo’s triumphs, a paper appeared that introduced the idea of using neural networks to represent quantum states, although it gave no indication of exactly how wide the tool’s reach might be. “We immediately recognized that this should be a very important paper,” Deng says, “so we put all our energy and time into studying the problem more.”

The result was a more complete account of the capabilities of certain neural networks to represent quantum states. In particular, the team studied neural networks that use two distinct groups of neurons. The first group, called the visible neurons, represents real quantum particles, like atoms in an optical lattice or ions in a chain. To account for interactions between particles, the researchers employed a second group of neurons—the hidden neurons—which link up with visible neurons. These links capture the physical interactions between real particles, and as long as the number of connections stays relatively small, the neural network description remains simple.

Specifying a number for each connection and mathematically forgetting the hidden neurons can produce a compact representation of many interesting quantum states, including states with topological characteristics and some with surprising amounts of entanglement.

Beyond its potential as a tool in numerical simulations, the new framework allowed Deng and collaborators to prove some mathematical facts about the families of quantum states represented by neural networks. For instance, neural networks with only short-range interactions—those in which each hidden neuron is only connected to a small cluster of visible neurons—have a strict limit on their total entanglement. This technical result, known as an area law, is a research pursuit of many condensed matter physicists.

These neural networks can’t capture everything, though. “They are a very restricted regime,” Deng says, adding that they don’t offer an efficient universal representation. If they did, they could be used to simulate a quantum computer with an ordinary computer, something physicists and computer scientists think is very unlikely. Still, the collection of states that they do represent efficiently, and the overlap of that collection with other representation methods, is an open problem that Deng says is ripe for further exploration.

By Chris Cesare, This email address is being protected from spambots. You need JavaScript enabled to view it. 

REFERENCE PUBLICATION
"Quantum Entanglement in Neural Network States," D.L. Deng, X. Li, D. Sarma, PHYSICAL REVIEW X, 7, (2017)
RESEARCH CONTACT
Dongling Deng | This email address is being protected from spambots. You need JavaScript enabled to view it.

Gravitational Waves Again Detected

LigoGW170104Dancing Duo of Black Holes (Credit: LIGO)

Scientists in the Laser Interferometer Gravitational-Wave Observatory (LIGO) collaboration now report the detection of a third gravitational wave event, named GW170104 and described in the July 1, 2017 Physical Review Letters. UMD's Peter Shawhan and Alessandra Buonanno are principal investigators in the experiment.

For a detailed account, see the CMNS story here.

 

Tiny tug unleashes cryogenic currents

A crystal of samarium hexaboride sits suspended between two titanium supports. (Credit: A. Stern/UCI)A crystal of samarium hexaboride sits suspended between two titanium supports. (Credit: A. Stern/UCI)

Researchers have found that a small stretch is enough to unleash the exotic electrical properties of a newly discovered topological insulator, unshackling a behavior previously locked away at cryogenic temperatures.

The compound, called samarium hexaboride, has been studied for decades. But recently it has enjoyed a surge of renewed interest as scientists first predicted and then discovered that it was a new type of topological insulator—a material that banishes electrical currents from its interior and forces them to travel along its periphery. That behavior only emerges at around 4 degrees above absolute zero, though, thwarting potential applications.

Now, experimentalists at the University of California, Irvine (UCI), working with JQI Fellow Victor Galitski and former JQI postdoctoral researcher Maxim Dzero (now at Kent State University), have found a way to activate samarium hexaboride's cryogenic behavior at much higher temperatures. By stretching small crystals of the metal by less than a percent, the team was able to spot the signature surface currents of a topological insulator at 240 K (minus 33 C)—nearly room temperature and, in any case, a far cry from 4 K. The currents even persisted once the strain was removed.

Their technique, which was recently reported in Nature Materials, uses piezoelectric elements that bend when they are fed with an electric current. By suspending a sample of samarium hexaboride between two titanium supports and pulling on one side, researchers could measure the crystal's electrical properties for different temperatures and amounts of stretch.

Last year, Galitski partnered with the same experimental group at UCI and discovered a potential application for samarium hexaboride's unusual surface currents. They found that holding a small crystal at a fixed voltage could produce oscillating currents on its surface. Such tick-tock signals are at the heart of modern digital electronics, but they typically require clocks that are much larger than the micron-sized crystals.

The new result might make such applications more likely, and it could even be achieved without any piezo elements. It may be possible to grow samarium hexaboride as a thin film on top of another material that would naturally cause it to stretch, the researchers say.

REFERENCE PUBLICATION
"Surface-dominated conduction up to 240K in the Kondo insulator SmB6 under strain," A. Stern, M. Dzero, V.M. Galitski, Z. Fisk, J. Xia, Nature Materials, advance online publication, – (2017)

RESEARCH CONTACT
This email address is being protected from spambots. You need JavaScript enabled to view it.

MEDIA CONTACT
This email address is being protected from spambots. You need JavaScript enabled to view it.

RELATED JQI ARTICLES
Oscillating currents point to practical application for topological insulators
An Ideal Material
Topological Insulators

Trapped ions and superconductors face off in quantum benchmark

debnath nature conceptAn artist's rendering of many linked trapped-ion modules. Researchers at JQI put one of their modules to the test against an IBM superconducting device. (Credit: E. Edwards/JQI)

The race to build larger and larger quantum computers is heating up, with several technologies competing for a role in future devices. Each potential platform has strengths and weaknesses, but little has been done to directly compare the performance of early prototypes. Now, researchers at the JQI have performed a first-of-its-kind benchmark test of two small quantum computers built from different technologies.

The team, working with JQI Fellow Christopher Monroe and led by postdoctoral researcher Norbert Linke, sized up their own small-scale quantum computer against a device built by IBM. Both machines use five qubits—the fundamental units of information in a quantum computer—and both machines have similar error rates. But while the JQI device relies on chains of trapped atomic ions, IBM Q uses coupled regions of superconducting material.

To make their comparison, the JQI team ran several quantum programs on the devices, each of which solved a simple problem using a series of logic gates to manipulate one or two qubits at a time. Researchers accessed the IBM device using an online interface, which allows anyone to try their hand at programming IBM Q.

Both computers have strengths and weaknesses. For example, the superconducting platform has quicker gates and may be easier to mass produce, but its man-made qubits are all slightly different and have shorter lifetimes. Monroe says that the slower gates of ions might not be a major hurdle, though. "Because there is time," Monroe says. "Trapped ion qubit lifetimes are way longer than any other type of qubit. Moreover, the ion qubits are identical, and they can be better replicated without error."

When put to the test, researchers found that the trapped-ion module was more accurate for programs that involved many pairs of qubits. Linke and Monroe attribute this to the simple fact that every qubit in their device is connected to every other—meaning that a logic gate can connect any pair of qubits. IBM Q has fewer than half the connections of its JQI counterpart, and in order to run some programs it had to shuffle information between qubits—a step that introduced errors into the calculation. When this shuffling wasn't necessary, the two computers had similar performance. "As we build larger systems, connectivity between qubits will become even more important," Monroe says.

The new study, which was recently published in Proceedings of the National Academy of Sciences, provides an important benchmark for researchers studying quantum computing. And such head-to-head comparisons will become increasingly important in the future. "If you want to buy a quantum computer, you'll need to know which one is best for your application," Linke says. "You'll need to test them in some way, and this is the first of this kind of comparison."

By Erin Marshall

REFERENCE PUBLICATION
"Experimental comparison of two quantum computing architectures," N.M. Linke, D. Maslov, M. Roetteler, S. Debnath, C. Figgatt, K.A. Landsman, K. Wright, C. Monroe, Proceedings of the National Academy of Sciences, 114, 3305-3310 (2017)
RESEARCH CONTACT
Norbert Linke|This email address is being protected from spambots. You need JavaScript enabled to view it.

Christopher Monroe|This email address is being protected from spambots. You need JavaScript enabled to view it.

RELATED JQI ARTICLES
Programmable ions set the stage for general-purpose quantum computers

The latest on HAWC and the search for high-energy gamma rays

In our own galaxy and beyond, violent collisions fling a never-ending stream of stuff at the earth, and astrophysicists are eager to learn more about the processes that produce this cosmic barrage.

Researchers from around the world have teamed up to build the High-Altitude Water Cherenkov (HAWC) gammy-ray observatory, an array of hundreds of huge water tanks on a mountain in Mexico. HAWC helps astrophysicists spot active cosmic neighborhoods by capturing the shower of particles created when high-energy packets of light smash into the earth's atmosphere.

Jordan Goodman, HAWC's lead investigator, and Dan Fiorino, a postdoctoral researcher at UMD, tell Chris Cesare about the details of the HAWC experiment and how it promises to fill some gaps in our understanding of the universe. To learn more about HAWC, please visit www.hawc-observatory.org. The collaboration is preparing to publish the first results of its search, and you can read about the details in an upcoming source catalog or a paper about high-energy gamma rays from the Crab Nebula.

This episode of Relatively Certain was produced by Chris Cesare, Sean Kelley and Emily Edwards and edited by Chris Cesare and Kate Delossantos, featuring music by Dave Depper, Podington Bear, Kevin MacLeod and Chris Zabriskie. Relatively Certain is a production of the Joint Quantum Institute and the University of Maryland, and you can find it on iTunes, Google Play or Soundcloud. 

To listen now, click here and a player will open.