Menu

Blog

Page 6418

Jan 12, 2021

Diffractive networks improve optical image classification accuracy

Posted by in categories: information science, robotics/AI

Recently, there has been a reemergence of interest in optical computing platforms for artificial intelligence-related applications. Optics is ideally suited for realizing neural network models because of the high speed, large bandwidth and high interconnectivity of optical information processing. Introduced by UCLA researchers, Diffractive Deep Neural Networks (D2NNs) constitute such an optical computing framework, comprising successive transmissive and/or reflective diffractive surfaces that can process input information through light-matter interaction. These surfaces are designed using standard deep learning techniques in a computer, which are then fabricated and assembled to build a physical optical network. Through experiments performed at terahertz wavelengths, the capability of D2NNs in classifying objects all-optically was demonstrated. In addition to object classification, the success of D2NNs in performing miscellaneous optical design and computation tasks, including e.g., spectral filtering, spectral information encoding, and optical pulse shaping have also been demonstrated.

In their latest paper published in Light: Science & Applications, UCLA team reports a leapfrog advance in D2NN-based image classification accuracy through ensemble learning. The key ingredient behind the success of their approach can be intuitively understood through the experiment of Sir Francis Galton (1822–1911), an English philosopher and statistician, who, while visiting a livestock fair, asked the participants to guess the weight of an ox. None of the hundreds of participants succeeded in guessing the weight. But to his astonishment, Galton found that the median of all the guesses came quite close—1207 pounds, and was accurate within 1% of the true weight of 1198 pounds. This experiment reveals the power of combining many predictions in order to obtain a much more accurate prediction. Ensemble learning manifests this idea in machine learning, where an improved predictive performance is attained by combining multiple models.

In their scheme, UCLA researchers reported an ensemble formed by multiple D2NNs operating in parallel, each of which is individually trained and diversified by optically filtering their inputs using a variety of filters. 1252 D2NNs, uniquely designed in this manner, formed the initial pool of networks, which was then pruned using an iterative pruning algorithm, so that the resulting physical ensemble is not prohibitively large. The final prediction comes from a weighted average of the decisions from all the constituent D2NNs in an ensemble. The researchers evaluated the performance of the resulting D2NN ensembles on CIFAR-10 image dataset, which contains 60000 natural images categorized in 10 classes and is an extensively used dataset for benchmarking various machine learning algorithms. Simulations of their designed ensemble systems revealed that diffractive optical networks can significantly benefit from the ‘wisdom of the crowd’.

Jan 12, 2021

Machine learning accelerates discovery of materials for use in industrial processes

Posted by in categories: materials, robotics/AI

New research led by researchers at the University of Toronto (U of T) and Northwestern University employs machine learning to craft the best building blocks in the assembly of framework materials for use in a targeted application.

Jan 12, 2021

Discovery of quantum behavior in insulators suggests possible new particle

Posted by in categories: particle physics, quantum physics

In a surprising discovery, Princeton physicists have observed an unexpected quantum behavior in an insulator made from a material called tungsten ditelluride. This phenomenon, known as quantum oscillation, is typically observed in metals rather than insulators, and its discovery offers new insights into our understanding of the quantum world. The findings also hint at the existence of an entirely new type of quantum particle.

The discovery challenges a long-held distinction between metals and insulators, because in the established quantum theory of materials, insulators were not thought to be able to experience quantum oscillations.

“If our interpretations are correct, we are seeing a fundamentally new form of quantum matter,” said Sanfeng Wu, assistant professor of physics at Princeton University and the senior author of a recent paper in Nature detailing this new discovery. “We are now imagining a wholly new quantum world hidden in insulators. It’s possible that we simply missed identifying them over the last several decades.”

Jan 12, 2021

What is Elon Musk’s Starship?

Posted by in categories: Elon Musk, space travel

Elon Musk’s company SpaceX is building a vehicle that could transform space travel.

Jan 12, 2021

Reality Does Not Depend on the Measurer According to New Interpretation of Quantum Mechanics

Posted by in categories: neuroscience, quantum physics

For 100 years scientists have disagreed on how to interpret quantum mechanics. A recent study by Jussi Lindgren and Jukka Liukkonen supports an interpretation that is close to classical scientific principles.

Quantum mechanics arose in the 1920s – and since then scientists have disagreed on how best to interpret it. Many interpretations, including the Copenhagen interpretation presented by Niels Bohr and Werner Heisenberg and in particular von Neumann-Wigner interpretation, state that the consciousness of the person conducting the test affects its result. On the other hand, Karl Popper and Albert Einstein thought that an objective reality exists. Erwin Schrödinger put forward the famous thought experiment involving the fate of an unfortunate cat that aimed to describe the imperfections of quantum mechanics.

Jan 12, 2021

Scientists Discover a Way to Control the Immune System’s “Natural Killer” Cells With “Invisible” Stem Cells

Posted by in categories: biotech/medical, business, life extension

UC San Francisco scientists have discovered a new way to control the immune system’s “natural killer” (NK) cells, a finding with implications for novel cell therapies and tissue implants that can evade immune rejection. The findings could also be used to enhance the ability of cancer immunotherapies to detect and destroy lurking tumors.

The study, published today (January 82021) in the Journal of Experimental Medicine, addresses a major challenge for the field of regenerative medicine, said lead author Tobias Deuse, MD, the Julien I.E. Hoffman, MD, Endowed Chair in Cardiac Surgery in the UCSF Department of Surgery.

“As a cardiac surgeon, I would love to put myself out of business by being able to implant healthy cardiac cells to repair heart disease,” said Deuse, who is interim chair and director of minimally invasive cardiac surgery in the Division of Adult Cardiothoracic Surgery. “And there are tremendous hopes to one day have the ability to implant insulin-producing cells in patients with diabetes or to inject cancer patients with immune cells engineered to seek and destroy tumors. The major obstacle is how to do this in a way that avoids immediate rejection by the immune system.”

Jan 11, 2021

Entangled photons can see through translucent materials

Posted by in categories: biological, quantum physics

Quantum twist on optical coherence tomography offers million-fold improvement in imaging.


Entangled pairs of photons have been used by physicists in Germany and Austria to image structures beneath the surfaces of materials that scatter light. The research was led by Aron Vanselow and Sven Ramelow at Humboldt University of Berlin and achieved high-resolution images of the samples using “ultra-broadband” photon pairs with very different wavelengths. One photon probed the sample, while the other read out image information. Their compact, low-cost and non-destructive system could be put to work inspecting advanced ceramics and mixing in fluids.

Optical coherence tomography (OCT) is a powerful tool for imaging structures beneath the surfaces of translucent materials and has a number of applications including the 3D scanning of biological tissues. The technique uses interferometry to reject the majority of light that has scattered many times in an object, focussing instead on the rare instances when light only scatters once from a feature of interest. This usually involves probing the material with visible or near-infrared light, which can be easily produced and detected. Yet in some materials such as ceramics, paints, and micro-porous samples, visible and near-infrared light is strongly scattered – which limits the use of OCT. Mid-infrared light, however, can penetrate deeper into these samples without scattering – but this light is far more difficult to produce and detect.

Continue reading “Entangled photons can see through translucent materials” »

Jan 11, 2021

It’s Not Your Rubber Tires That Protect You From Lightning

Posted by in category: climatology

Circa 2016


Many people think that it is the rubber tires that protect them when their car is struck by lightning. In reality, their car is becoming a Faraday cage. What is that and how does it work?

Michael Faraday was a British scientist born in 1791. Although not formally educated, he had a strong interest in electromagnetism. He also credited with discovering Benzene and popularizing terms such as anode, cathode and electrode. As an apprentice for a bookbinder, he read many books which encouraged his interest in science. He soon became a well known experimental scientist leading to his name becoming a unit of electrical charge. He is also known for inventing the Faraday rotator and Faraday cage.

Jan 11, 2021

CRISPR gene editing used to store data in DNA inside living cells

Posted by in categories: bioengineering, biotech/medical

Biologists have used CRISPR gene editing to store information inside DNA in living bacterial cells, which could become a storage medium of the future.

Jan 11, 2021

Largest spider

Posted by in category: futurism