System Bits: March 27

New quantum electronic material has atomic structure resembling a Japanese basketweaving pattern
According to MIT, Harvard University, and Lawrence Berkeley National Laboratory researchers, a motif of Japanese basketweaving known as the kagome pattern has preoccupied physicists for decades. They reminded that kagome baskets are typically made from strips of bamboo woven into a highly symmetrical pattern of interlaced, corner-sharing triangles. And if a metal or other conductive material could be made to resemble such a kagome pattern at the atomic scale, with individual atoms arranged in similar triangular patterns, it should in theory exhibit exotic electronic properties, they said.

Now, the team said they’ve produced a kagome metal, which is an electrically conducting crystal, made from layers of iron and tin atoms, with each atomic layer arranged in the repeating pattern of a kagome lattice.

An illustration depicting a kagome metal — an electrically conducting crystal, made from layers of iron and tin atoms, with each atomic layer arranged in the repeating pattern of a kagome lattice.
Source: MIT

When the researchers flowed a current across the kagome layers within the crystal, they observed that the triangular arrangement of atoms induced strange, quantum-like behaviors in the passing current. Instead of flowing straight through the lattice, electrons instead veered, or bent back within the lattice.

This behavior is a 3D cousin of the so-called Quantum Hall effect, in which electrons flowing through a 2D material will exhibit a chiral, topological state, in which they bend into tight, circular paths and flow along edges without losing energy.

Joseph Checkelsky, assistant professor of physics at MIT explained, “By constructing the kagome network of iron, which is inherently magnetic, this exotic behavior persists to room temperature and higher. The charges in the crystal feel not only the magnetic fields from these atoms, but also a purely quantum-mechanical magnetic force from the lattice. This could lead to perfect conduction, akin to superconductivity, in future generations of materials.”

To explore these findings, the team said they measured the energy spectrum within the crystal, using a modern version of an effect first discovered by Heinrich Hertz and explained by Einstein, known as the photoelectric effect. 

The team is now investigating ways to stabilize other more highly 2D kagome lattice structures. Such materials, if they can be synthesized, could be used to explore not only devices with zero energy loss, such as dissipationless power lines, but also applications toward quantum computing.

2D quantum bits
A team led by Technische Universität Wien has performed theoretical simulations proving that two novel materials, each composed of a single atomic layer and the tip of a scanning tunneling microscope – are the ingredients to create a novel kind of a so-called quantum dot. These extremely small nanostructures allow delicate control of individual electrons by fine-tuning their energy levels directly, and are key for modern quantum technologies.

New kinds of quantum bits: Tiny nanostructures allow delicate control of individual electrons.
Source: Technische Universität Wien

Many applications in the field of quantum technologies require a quantum system in which electrons occupy two states – similar to a classical switch – on or off, with the difference that quantum physics also allows for arbitrary superpositions of the on and off states, explained Florian Libisch from the Institute for Theoretical Physics at Technische Universität Wien.

A key property of such systems is the energy difference between those two quantum states such that efficiently manipulating the information stored in the quantum state of the electrons requires perfect control of the system parameters, Libisch said. “An ideal system allows for continuous tuning the energy difference from zero to a large value.”

For systems found in nature, such as atoms, this is usually difficult to realize because the energies of atomic states, and hence their differences, are fixed. Tuning energies becomes possible in synthetic nanostructures engineered towards confining electrons. Such structures are often referred to as quantum dots or artificial atoms.

The research team of TU Wien, RWTH Aachen and the University of Manchester now succeeded in developing a new type of quantum dot to allow for much more accurate and widely tunable energy levels of confined electrons. They achieved this by combining two very special materials: graphene, a conductive single atomic layer of carbon atoms, and hexagonal boron nitride, also a single layer of material quite similar to graphene except that it is insulating.

As a next step, the tip of the scanning tunneling microscope could be replaced by a series of nanoelectronic gates, which would allow for exploiting the quantum dot states of graphene on hexagonal boron nitride for scalable quantum technologies such as “valleytronics.”

Evaluating AI’s applicability
As AI’s role in society continues to expand, J. B. Brown of the Kyoto University Graduate School of Medicine reports on a new evaluation method for the type of AI that predicts yes/positive/true or no/negative/false answers.

Brown’s paper deconstructs the utilization of AI and analyzes the nature of the statistics used to report an AI program’s ability. The new technique also generates a probability of the performance level given evaluation data, answering questions such as: What is the probability of achieving accuracy greater than 90%?

The new AI evaluation method looks at the input data itself to find it the ‘accuracy’ of the AI can be trusted
Source: Kyoto University

Reports of new AI applications appear in the news almost daily, including in society and science, finance, pharmaceuticals, medicine, and security, and while reported statistics seem impressive, research teams and those evaluating the results come across two problems, Brown said.

First, to understand if the AI achieved its results by chance, and second, to interpret applicability from the reported performance statistics.

For example, if an AI program is built to predict whether or not someone will win the lottery, it may always predict a loss. The program may achieve 99% accuracy, but interpretation is key to determine the accuracy of the conclusion that the program is accurate.

Herein lies the problem, Brown said. In typical AI development, the evaluation can only be trusted if there is an equal number of positive and negative results. If the data is biased toward either value, the current system of evaluation will exaggerate the system’s ability. To tackle this problem, Brown developed a new technique that evaluates performance based on only the input data itself. “The novelty of this technique is that it doesn’t depend on any one type of AI technology, such as deep learning. It can help develop new evaluation metrics by looking at how a metric interplays with the balance in predicted data. We can then tell if the resulting metrics could be biased.”

Brown hopes this analysis will not only raise awareness of how we think about AI in the future, but also that it contributes to the development of more robust AI platforms.