Neural networks are taking the world of computing by storm. Researchers have used them to create machines that are learning a huge range of skills that had previously been the unique preserve of humans--object recognition, face recognition, natural language processing, machine translation. All these skills, and more, are now becoming routine for machines. So there is great interest in creating more capable neural networks that can push the boundaries of artificial intelligence even further. The focus of this work is in creating circuits that operate more like neurons, so-called neuromorphic chips.
Researchers at the National Institute of Standards and Technology (NIST) have made a silicon chip that distributes optical signals precisely across a miniature brain-like grid, showcasing a potential new design for neural networks. The human brain has billions of neurons (nerve cells), each with thousands of connections to other neurons. Many computing research projects aim to emulate the brain by creating circuits of artificial neural networks. But conventional electronics, including the electrical wiring of semiconductor circuits, often impedes the extremely complex routing required for useful neural networks. The NIST team proposes to use light instead of electricity as a signaling medium.
MIT researchers have developed a novel "photonic" chip that uses light instead of electricity -- and consumes relatively little power in the process. The chip could be used to process massive neural networks millions of times more efficiently than today's classical computers do. Neural networks are machine-learning models that are widely used for such tasks as robotic object identification, natural language processing, drug development, medical imaging, and powering driverless cars. Novel optical neural networks, which use optical phenomena to accelerate computation, can run much faster and more efficiently than their electrical counterparts. But as traditional and optical neural networks grow more complex, they eat up tons of power.
Many complex systems operate with loss. Mathematically, these systems can be described as non-Hermitian. A property of such a system is that there can exist certain conditions--exceptional points--where gain and loss can be perfectly balanced and exotic behavior is predicted to occur. Optical systems generally possess gain and loss and so are ideal systems for exploring exceptional point physics. Miri and Alù review the topic of exceptional points in photonics and explore some of the possible exotic behavior that might be expected from engineering such systems. Singularities are critical points for which the behavior of a mathematical model governing a physical system is of a fundamentally different nature compared to the neighboring points. Exceptional points are spectral singularities in the parameter space of a system in which two or more eigenvalues, and their corresponding eigenvectors, simultaneously coalesce. Such degeneracies are peculiar features of nonconservative systems that exchange energy with their surrounding environment. In the past two decades, there has been a growing interest in investigating such nonconservative systems, particularly in connection with the quantum mechanics notions of parity-time symmetry, after the realization that some non-Hermitian Hamiltonians exhibit entirely real spectra. Lately, non-Hermitian systems have raised considerable attention in photonics, given that optical gain and loss can be integrated as nonconservative ingredients to create artificial materials and structures with altogether new optical properties. As we introduce gain and loss in a nanophotonic system, the emergence of exceptional point singularities dramatically alters the overall response, leading to a range of exotic functionalities associated with abrupt phase transitions in the eigenvalue spectrum. Even though such a peculiar effect has been known theoretically for several years, its controllable realization has not been made possible until recently and with advances in exploiting gain and loss in guided-wave photonic systems. As shown in a range of recent theoretical and experimental works, this property creates opportunities for ultrasensitive measurements and for manipulating the modal content of multimode lasers. In addition, adiabatic parametric evolution around exceptional points provides interesting schemes for topological energy transfer and designing mode and polarization converters in photonics.
Deep learning is having a serious moment right now in the world of AI. Loosely based on the brain's computing architecture, artificial neural networks have vastly outperformed their predecessors in a variety of tasks that had previously stumped our silicon-minded comrades. But as these algorithms continuously forge new grounds in machine intelligence, we're coming to an uncomfortable realization: transistor-based computers have hard limits, and those limits are approaching rapidly. Now, thanks to a new system developed by Princeton engineers, we may have one way to smash the speed barrier of our current processors: neuromorphic computing running on photons, not electrons, with silicon chips that work at the speed of light. Published this week on Arxiv, the new photonic neural network is so blazingly efficient that when pitted against a conventional CPU in solving differential equations, it performed roughly 2,000 times faster.