Directional routing of single photons

Amazing stuff! However, through Google search I found what appears to be a  similar work published in 2015 (a PhD dissertation).

"Photons are robust and can travel long distances, making them ideal carriers of quantum information. However, communication between nodes on a network requires directional control of the photons. Emission from excited atoms, for instance, can generally take any direction.

Li et al. demonstrated a direction-switchable single-photon emitter using Rydberg polaritons. The ensemble of cesium atoms is first excited with a laser pulse, and then a second retrieval pulse is used to de-excite the system and extract the stored photon. The direction of the retrieval laser determines the direction of the emitted photon. Using this protocol, the angle difference between the incoming and redirected photon can be up to 100°. Such control over the directional emission should prove useful for quantum communication."

From the abstract:

"A promising route toward quantum networking is via photons as information carriers, requiring deterministic quantum nonlinear optical operations and single-photon routing.

Here, we demonstrate a direction-switchable single-photon emitter using a Rydberg polariton. The Rydberg component of the stored photon is changed using a stimulated Raman transition with a specific intermediate state.

By adjusting the direction of the retrieval laser, we can redirect the emitted photon into a rich variety of alternative modes. We experimentally demonstrate a redirection angle of  . 

Building upon this scheme, we propose a quantum routing of single photons with  output channels by rotation of the retrieval laser, where all directions have identical routing efficiency. In addition, the protocol reduces the effect of motional dephasing, increasing the photon lifetime to µs (  times the photon processing time), enabling functional quantum devices based on Rydberg polaritons."

In Other Journals | Science

Direction-switchable single-photon emitter using a Rydberg polariton (open access)

Fig. 1. Experimental realization and relevant energy levels. 

Photonics advance could enable compact, high-performance lidar sensors with no moving parts

Good news!

"... A new study from ... researchers could help to enable next-generation lidar sensors that are compact, durable, and have no moving parts. The key advance is a novel design for a silicon-photonics chip, which is a semiconductor device that manipulates light rather than electricity.  ...

To avoid these drawbacks, the ... researchers designed and demonstrated an array of integrated antennas that minimizes unwanted crosstalk between the antennas. Their innovation allows a lidar chip to scan a wider field of view while maintaining low-noise operation compared to other silicon-photonics-based approaches. ..."

From the abstract:

"Integrated optical phased arrays (OPAs) have emerged as a promising technology for many applications due to their ability to dynamically control free-space optical beams in a compact and non-mechanical manner.

However, these integrated OPAs typically have a restricted field of view (FOV), limited by grating lobes caused by large antenna pitches that are typically necessary to reduce crosstalk between the antennas in the integrated OPA.

In this work, we develop and experimentally demonstrate for the first time, to the best of our knowledge, a set of integrated grating-based antennas with significantly-reduced inter-antenna crosstalk that enable half-wavelength-pitch integrated OPAs with grating-lobe-free and wide-FOV functionality.

First, we derive a generalized theoretical model to describe the coupling dynamics between lossy modes in a system and use this model to analyze the coupling between antennas.

Next, we design and demonstrate a set of three integrated grating-based antennas with different propagation coefficients to enable reduced inter-antenna crosstalk, successfully measuring a significant reduction from 100% to 1% coupling.

Finally, using these reduced-crosstalk antennas, we develop and demonstrate a half-wavelength-pitch integrated OPA, successfully demonstrating grating-lobe-free and wide-FOV functionality.

This work facilitates new functionality for high-performance integrated OPAs."

Photonics advance could enable compact, high-performance lidar sensors | MIT News | Massachusetts Institute of Technology "With a novel design, MIT researchers overcame a stubborn problem that has limited the effectiveness of chip-based systems for lidar."

Reduced-crosstalk antennas for grating-lobe-free and wide-field-of-view integrated optical phased arrays (open access)

Fig. 1: Wide-FOV integrated-OPA concept.

Fig. 2: Design of reduced-crosstalk antennas.

Jelena Notaros, senior author (Source)

New photonic chip boosts AI efficiency with light to cut energy use

Good news!

"... To tackle the growing energy demands of AI, researchers ... have built something dazzling. Their new chip, called a photonic joint transform correlator (pJTC), swaps electricity for light to handle one of AI's most power-hungry jobs. ..."

"... The prototype chip uses two sets of miniature Fresnel lenses using standard manufacturing processes. These two-dimensional versions of the same lenses found in lighthouses are just a fraction of the width of a human hair. Machine learning data, such as from an image or other pattern-recognition tasks, are converted into laser light on-chip and passed through the lenses. The results are then converted back into a digital signal to complete the AI task.

This lens-based convolution system is not only more computationally efficient, but it also reduces the computing time. Using light instead of electricity has other benefits, too. Sorger’s group designed a chip that could use different colored lasers to process multiple data streams in parallel. ..."

From the abstract:

"Convolutional operations are computationally intensive in artificial intelligence services, and their overhead in electronic hardware limits machine learning scaling. Here, we introduce a photonic joint transform correlator (pJTC) using a near-energy-free on-chip Fourier transformation to accelerate convolution operations. The pJTC reduces computational complexity for both convolution and cross-correlation from O(N4) to O(N2), where N2 is the input data size.

Demonstrating functional Fourier transforms and convolution, this pJTC achieves 98.0% accuracy on an exemplary MNIST inference task.

Furthermore, a wavelength-multiplexed pJTC architecture shows potential for high throughput and energy efficiency, reaching 305 TOPS/W and 40.2 TOPS/mm2, based on currently available foundry processes.

An efficient, compact, and low-latency convolution accelerator promises to advance next-generation AI capabilities across edge demands, high-performance computing, and cloud services."

New photonic chip boosts AI efficiency with light

New light-based chip boosts power efficiency of AI tasks 100 fold (original news release)

New light-based chip boosts power efficiency of AI tasks 100 fold (original news release)

Near-energy-free Photonic Fourier Transformation for Convolution Operation Acceler (preprint, open access)

pJTC Convolutional techniques.

(a) Comparison of spatial convolution, Fourier electrical convolution, and Fourier optical convolution in terms of computational complexity.

(b) Schematic of a JTC, demonstrating how it performs Fourier optical convolution by optically generating the Fourier transform of the combined input Signal and Kernel, detecting the intensity pattern, and producing the auto- and cross-correlation between Signal and Kernel.

(c) Optical microscope image of the fabricated SiPh chiplet from AIM Photonics. (d) Comparison of the initial MNIST image (green line), the output after an ideal Fourier transform (blue line), the output after the actual on-chip lens Fourier transform (yellow line), and the calibrated output obtained from the actual on-chip lens after applying phase correction (pink line). (e) Confusion matrix shows the classification accuracy for 10,000 test MNIST images with 10 percent random temporal delay introduced in the input electrical signal, achieving total accuracy of 95.3 percent.

Ultrasmall optical devices rewrite the rules of light manipulation

Good news!

"In the push to shrink and enhance technologies that control light, MIT researchers have unveiled a new platform that pushes the limits of modern optics through nanophotonics, the manipulation of light on the nanoscale, or billionths of a meter.

The result is a class of ultracompact optical devices that are not only smaller and more efficient than existing technologies, but also dynamically tunable, or switchable, from one optical mode to another. Until now, this has been an elusive combination in nanophotonics. ...

CrSBr is a layered quantum material with a rare combination of magnetic order and strong optical response. Central to its unique optical properties are excitons: quasiparticles formed when a material absorbs light and an electron is excited, leaving behind a positively charged “hole.” The electron and hole remain bound together by electrostatic attraction, forming a sort of neutral particle that can strongly interact with light.

In CrSBr, excitons dominate the optical response and are highly sensitive to magnetic fields, which means they can be manipulated using external controls. ...

Because of these excitons, CrSBr exhibits an exceptionally large refractive index that allows researchers to sculpt the material to fabricate optical structures like photonic crystals that are up to an order of magnitude thinner than those made from traditional materials. “We can make optical structures as thin as 6 nanometers, or just seven layers of atoms stacked on top of each other,” ...

And crucially, by applying a modest magnetic field, the MIT researchers were able to continuously and reversibly switch the optical mode. In other words, they demonstrated the ability to dynamically change how light flows through the nanostructure, all without any moving parts or changes in temperature. ..."

From the abstract:

"Central to the field of nanophotonics is the ability to engineer the flow of light through nanoscale structures. These structures often have permanent working spectral ranges and optical properties that are fixed during fabrication.

Quantum materials, with their correlated and intertwined degrees of freedom, offer a promising avenue for dynamically controlling photonic devices without altering their physical structure.

Here we fabricate photonic crystal slabs from CrSBr, a van der Waals antiferromagnetic semiconductor, and demonstrate in situ control over their optical properties.

Leveraging the combination of the exceptionally large refractive index of CrSBr near its excitonic resonances and its tunability via external fields, we achieve precise manipulation of photonic modes at near-visible and infrared wavelengths, showcasing a new paradigm for nanophotonic device design.

The resulting guided resonances of the photonic crystal are tightly packed in the spectrum with very small mode volumes, are highly tunable via external magnetic fields and exhibit high Q factors exceeding 1,200.

These resonances self-hybridize with the excitonic degrees of freedom, resulting in intrinsic strong light–matter coupling.

Our findings underscore the potential of quantum materials for developing in situ tunable photonic elements and cavities."

Ultrasmall optical devices rewrite the rules of light manipulation | MIT News | Massachusetts Institute of Technology "Nanophotonic devices developed at MIT are compact, efficient, reprogrammable, adaptive, and able to dynamically respond to external inputs."

Tunable nanophotonic devices and cavities based on a two-dimensional magnet (no public access)

Tunable Nanophotonic Devices and Cavities based on a Two-Dimensional Magnet (preprint, open access)

Photonic computer chips perform as well as purely electronic counterparts, but faster

Good news! Amazing stuff!

"Researchers in Singapore and the US have independently developed two new types of photonic computer chips that match existing purely electronic chips in terms of their raw performance. The chips, which can be integrated with conventional silicon electronics, could find use in energy-hungry technologies such as artificial intelligence (AI). ...

Light-based computation, which exploits photons instead of electrons, is a promising alternative because it can perform multiplication and accumulation (MAC) much more quickly and efficiently than electronic devices. ...

The Singapore device was made by researchers at the photonic computing firm Lightelligence and is called PACE, for Photonic Arithmetic Computing Engine. It is a hybrid photonic-electronic system made up of more than 16 000 photonic components integrated on a single silicon chip and performs matrix MAC on 64-entry binary vectors. ...

The Lightelligence device, which the team describe in Nature, can solve complex computational problems known as max-cut/optimization problems that are important for applications in areas such as logistics. Notably, its greatly reduced minimum latency – a key measure of computation speed – means it can solve a type of problem known as an Ising model in just five nanoseconds. This makes it 500 times faster than today’s best graphical-processing-unit-based systems at this task. ...

Independently, researchers led by Nicholas Harris at Lightmatter in Mountain View, California, have fabricated the first photonic processor capable of executing state-of-the-art neural network tasks such as classification, segmentation and running reinforcement learning algorithms.

Lightmatter’s design consists of six chips in a single package with high-speed interconnects between vertically aligned photonic tensor cores (PTCs) and control dies. The team’s processor integrates four 128 x 128 PTCs, with each PTC occupying an area of 14 x 24.96 mm. It contains all the photonic components and analogue mixed-signal circuits required to operate and members of the team say that the current architecture could be scaled to 512 x 512 computing units in a single die.

The result is a device that can perform 65.5 trillion adaptive block floating-point 35 (ABFP) 16-bit operations per second with just 78 W of electrical power and 1.6 W of optical power. Writing in Nature, the researchers claim that this represents the highest level of integration achieved in photonic processing. ...

can implement complex AI models such as the neural network ResNet (used for image processing) and the natural language processing model BERT (short for Bidirectional Encoder Representations from Transformers) – all with an accuracy rivalling that of standard electronic processors. It can also compute reinforcement learning algorithms such as DeepMind’s Atari. ...

Both teams fabricated their photonic and electronic chips using standard complementary metal-oxide-semiconductor (CMOS) processing techniques. This means that existing infrastructures could be exploited to scale up their manufacture. Another advantage: both systems were fully integrated in a standard chip interface – a first. ..."

"... Computing stands at an inflection point unlike anything we’ve seen since the transistor was invented. Artificial intelligence workloads are driving computational demands beyond what traditional scaling laws—Moore’s Law, Dennard scaling, and memory scaling—can deliver. All three have effectively stalled, particularly on a per-silicon-area basis. ..."

From the abstract (1):

"Integrated photonics, particularly silicon photonics, have emerged as cutting-edge technology driven by promising applications such as short-reach communications, autonomous driving, biosensing and photonic computing. As advances in AI lead to growing computing demands, photonic computing has gained considerable attention as an appealing candidate. Nonetheless, there are substantial technical challenges in the scaling up of integrated photonics systems to realize these advantages, such as ensuring consistent performance gains in upscaled integrated device clusters, establishing standard designs and verification processes for complex circuits, as well as packaging large-scale systems. These obstacles arise primarily because of the relative immaturity of integrated photonics manufacturing and the scarcity of advanced packaging solutions involving photonics.

Here we report a large-scale integrated photonic accelerator comprising more than 16,000 photonic components. The accelerator is designed to deliver standard linear matrix multiply–accumulate (MAC) functions, enabling computing with high speed up to 1 GHz frequency and low latency as small as 3 ns per cycle. Logic, memory and control functions that support photonic matrix MAC operations were designed into a cointegrated electronics chip.

To seamlessly integrate the electronics and photonics chips at the commercial scale, we have made use of an innovative 2.5D hybrid advanced packaging approach. Through the development of this accelerator system, we demonstrate an ultralow computation latency for heuristic solvers of computationally hard Ising problems whose performance greatly relies on the computing latency."

From the abstract (2):

"Over the past decade, photonics research has explored accelerated tensor operations, foundational to artificial intelligence (AI) and deep learning, as a path towards enhanced energy efficiency and performance. The field is centrally motivated by finding alternative technologies to extend computational progress in a post-Moore’s law and Dennard scaling era.

Despite these advances, no photonic chip has achieved the precision necessary for practical AI applications, and demonstrations have been limited to simplified benchmark tasks.

Here we introduce a photonic AI processor that executes advanced AI models, including ResNet3 and BERT, along with the Atari deep reinforcement learning algorithm originally demonstrated by DeepMind. This processor achieves near-electronic precision for many workloads, marking a notable entry for photonic computing into competition with established electronic AI accelerators and an essential step towards developing post-transistor computing technologies."

Photonic computer chips perform as well as purely electronic counterparts, say researchers – Physics World

A New Kind of Computer (original press release)

An integrated large-scale photonic accelerator with ultralow latency (1, open access)

Universal photonic artificial intelligence acceleration (2, no public access)

Fig. 2: PACE system implementation.

Photonic processor PCI-e card top view and side view as well as a bottom view of the photonic processor chip package.

The Last Missing Piece of Silicon Photonics

Good news! This could be a breakthrough!

"International research team presents first electrically pumped continuous-wave semiconductor laser suitable for seamless silicon integration

Scientists from Forschungszentrum Jülich, FZJ, the University of Stuttgart, and the Leibniz Institute for High Performance Microelectronics (IHP), together with their French partner CEA-Leti, have developed the first electrically pumped continuous-wave semiconductor laser composed exclusively of elements from the fourth group of the periodic table – the “silicon group”. Built from stacked ultrathin layers of silicon germanium-tin and germanium-tin, this new laser is the first of its kind directly grown on a silicon wafer, opening up new possibilities for on-chip integrated photonics. ...

Key components, including high-performance modulators, photodetectors, and waveguides have been developed. However, a long-standing challenge has been the lack of an efficient, electrically pumped light source using only Group IV semiconductors. Until now, such light sources have traditionally relied on III-V materials, which are difficult and therefore expensive to integrate with silicon. ...

For the first time, the researchers have demonstrated continuous-wave operation in an electrically pumped Group IV laser on silicon. Unlike previous germanium-tin lasers that relied on high-energy optical pumping, this new laser operates with a low current injection of just 5 milliamperes (mA) at 2 volts (V), comparable to the energy consumption of a light-emitting diode. With its advanced multi-quantum well structure and ring geometry, the laser minimizes the power consumption and the heat generation, enabling stable operation up to 90 Kelvin (K) or minus 183.15 degrees Celsius (°C). ...

further optimizations are needed to further reduce the lasing threshold and achieve room-temperature operation. However, the success of earlier optically pumped germanium-tin lasers, which have evolved from cryogenic to room-temperature operation in only few years, suggests a clear path forward. ..."

From the abstract:

"Over the last 30 years, group-IV semiconductors have been intensely investigated in the quest for a fundamental direct bandgap semiconductor that could yield the last missing piece of the Si Photonics toolbox: a continuous-wave Si-based laser. Along this path, it has been demonstrated that the electronic band structure of the GeSn/SiGeSn heterostructures can be tuned into a direct bandgap quantum structure providing optical gain for lasing. In this paper, we present a versatile electrically pumped, continuous-wave laser emitting at a near-infrared wavelength of 2.32 µm with a low threshold current of 4 mA. It is based on a 6-periods SiGeSn/GeSn multiple quantum-well heterostructure. Operation of the micro-disk laser at liquid nitrogen temperature is possible by changing to pulsed operation and reducing the heat load. The demonstration of a continuous-wave, electrically pumped, all-group-IV laser is a major breakthrough towards a complete group-IV photonics technology platform."

Continuous-wave electrically pumped multi-quantum-well laser based on group-IV semiconductors (open access)

Credits: Neuer Halbleiterlaser löst zentrales Problem der Silizium-Photonik "Erster elektrisch gepumpter Laser für Siliziumchips entwickelt: Neue Lichtquelle aus Silizium-Germanium-Zinn optimiert die On-Chip-Photonik."

Fig. 1: SiGeSn/GeSn multi-quantum-well structure.

Scanning electron micrograph

Researchers demonstrate the first chip-based 3D printer

Amazing stuff! Is there something you can not put on a chip? 😊

"... Imagine a portable 3D printer you could hold in the palm of your hand. The tiny device could enable a user to rapidly create customized, low-cost objects on the go ...

Their proof-of-concept device consists of a single, millimeter-scale photonic chip that emits reconfigurable beams of light into a well of resin that cures into a solid shape when light strikes it.

The prototype chip has no moving parts, instead relying on an array of tiny optical antennas to steer a beam of light. The beam projects up into a liquid resin that has been designed to rapidly cure when exposed to the beam’s wavelength of visible light. ...

In the [future], they envision a system where a photonic chip sits at the bottom of a well of resin and emits a 3D hologram of visible light, rapidly curing an entire object in a single step. ..."

From the abstract:

"Imagine if it were possible to create 3D objects in the palm of your hand within seconds using only a single photonic chip. Although 3D printing has revolutionized the way we create in nearly every aspect of modern society, current 3D printers rely on large and complex mechanical systems to enable layer-by-layer addition of material. This limits print speed, resolution, portability, form factor, and material complexity.
Although there have been recent efforts in developing novel photocuring-based 3D printers that utilize light to transform matter from liquid resins to solid objects using advanced methods, they remain reliant on bulky and complex mechanical systems.
To address these limitations, we combine the fields of silicon photonics and photochemistry to propose the first chip-based 3D printer. The proposed system consists of only a single millimeter-scale photonic chip without any moving parts that emits reconfigurable visible-light holograms up into a simple stationary resin well to enable non-mechanical 3D printing. Furthermore, we experimentally demonstrate a stereolithography-inspired proof-of-concept version of the chip-based 3D printer using a visible-light beam-steering integrated optical phased array and visible-light-curable resin, showing 3D printing using a chip-based system for the first time. This work demonstrates the first steps towards a highly-compact, portable, and low-cost solution for the next generation of 3D printers."

Researchers demonstrate the first chip-based 3D printer | MIT News | Massachusetts Institute of Technology Smaller than a coin, this optical device could enable rapid prototyping on the go.

Silicon-photonics-enabled chip-based 3D printer (open access)

Fig. 1: The chip-based 3D printer concept.

Fig. 2: The 3D-printer integrated optical phased array architecture

Faster, More Secure Photonic Chip Boosts AI Training

Good news!

"A microchip that uses light instead of electricity can potentially be faster and more energy efficient at the complex computations essential to training AI than conventional electronics. In addition, researchers say the new chips may be significantly more secure against hacking. ...
“It might be around 1,000 to 10,000 times faster,” ...

In the new study, researchers created a silicon wafer that varied in height from 150 to 220 nanometers. The height variations were organized so that the chip could scatter light in specific patterns. When input in the form of light flows into the chip, the output light encodes data from complex tasks.

The scientists designed the microchip to perform vector matrix multiplication operations. These calculations, which involve multiplying grids of numbers known as matrices, are key to many computational tasks, including operating neural networks. ..."

From the abstract:

"Inverse-designed silicon photonic metastructures offer an efficient platform to perform analogue computations with electromagnetic waves. However, due to computational difficulties, scaling up these metastructures to handle a large number of data channels is not trivial. Furthermore, a typical inverse-design procedure is limited to a small computational domain and therefore tends to employ resonant features to achieve its objectives. This results in structures that are narrow-bandwidth and highly sensitive to fabrication errors. Here we employ a two-dimensional (2D) inverse-design method based on the effective index approximation with a low-index contrast constraint. This results in compact amorphous lens systems that are generally feed-forward and low-resonance. We designed and experimentally demonstrated a vector–matrix product for a 2 × 2 matrix and a 3 × 3 matrix. We also designed a 10 × 10 matrix using the proposed 2D computational method. These examples demonstrate that these techniques have the potential to enable larger-scale wave-based analogue computing platforms."

Faster, More Secure Photonic Chip Boosts AI Training - IEEE Spectrum Optical computing can perform matrix computations at the speed of light

Inverse-designed low-index-contrast structures on a silicon photonics platform for vector–matrix multiplication (no public access)

The area highlighted in red on this silicon photonic chip can perform 3-by-3-matrix operations, relevant for AI-related computations.

Breakthrough in computer chip energy efficiency could significantly cut data center and supercomputer electricity use

Good news! Sounds almost spectacular!

"... new, ultra-energy-efficient method to compensate for temperature variations that degrade photonic chips. Such chips “will form the high-speed communication backbone of future data centers and supercomputers,” ...

The issue with photonic chips is that up until now, significant energy has been required to keep their temperature stable and performance high. The team led by Wang, however, has shown that it’s possible to reduce the energy needed for temperature control by a factor of more than 1 million. ..."

From the abstract:

"Silicon microring resonators (Si-MRRs) play essential roles in on-chip wavelength division multiplexing (WDM) systems due to their ultra-compact size and low energy consumption. However, the resonant wavelength of Si-MRRs is very sensitive to temperature fluctuations and fabrication process variation. Typically, each Si-MRR in the WDM system requires precise wavelength control by free carrier injection using PIN diodes or thermal heaters that consume high power. This work experimentally demonstrates gate-tuning on-chip WDM filters for the first time with large wavelength coverage for the entire channel spacing using a Si-MRR array driven by high mobility titanium-doped indium oxide (ITiO) gates. The integrated Si-MRRs achieve unprecedented wavelength tunability up to 589 pm/V, or VπL of 0.050 V cm with a high-quality factor of 5200. The on-chip WDM filters, which consist of four cascaded ITiO-driven Si-MRRs, can be continuously tuned across the 1543–1548 nm wavelength range by gate biases with near-zero power consumption."

Breakthrough in computer chip energy efficiency could cut data center electricity use (secondary source)

Breakthrough in computer chip energy efficiency could cut data center electricity use (primary source)

On-chip wavelength division multiplexing filters using extremely efficient gate-driven silicon microring resonator array (open access)

Fig. 5 (a) Optical microscope image of the fabricated on-chip WDM filters consisting of four cascaded tunable Si-MRRs and testing setup (b) Zoom-in view of the individual tunable Si-MRR of the on-chip WDM filters. The dashed line highlights the ITiO gate. (c) The simulated carrier concentration (Nc), refractive index (n), and extinction coefficient (k) distributions with different applied biases at the ITiO/HfO2 and the Si/HfO2 interfaces.

Photonic Chips achieves first-ever optical backpropagation milestone

Good news! Possibly a breakthrough!

"Processors that use light instead of electricity show promise as a faster and more energy-efficient way to implement AI. So far they’ve only been used to run models that have already been trained, but new research has demonstrated the ability to train AI on an optical chip for the first time. ...
They are particularly promising for running AI because they are very efficient at carrying out matrix multiplications—a key calculation at the heart of all deep-learning models. ...
What sets the new chip apart though, is that it also has light sources and light detectors at both ends, allowing signals to pass forward and backward through the network. It also features small “taps” at each node in the network that siphon off a small amount of the light signal, redirecting it to an infrared camera that measures light intensities. Together, these changes make it possible to implement the optical backpropagation algorithm. The researchers showed that they could train a simple neural network to label points on a graph based on their position with an accuracy of up to 98 percent, which is comparable to conventional approaches. ..."

From the editorial summary and abstract:

"Editor’s summary

Commercial applications of machine learning (ML) are associated with exponentially increasing energy costs, requiring the development of energy-efficient analog alternatives. Many conventional ML methods use digital backpropagation for neural network training, which is a computationally expensive task. Pai et al. designed a photonic neural network chip to allow efficient and feasible in situ backpropagation training by monitoring optical power passing either forward or backward through each waveguide segment of the chip ... The presented proof-of-principle experimental realization of on-chip backpropagation training demonstrates one of the ways that ML could fundamentally change in the future, with most of the computation taking place optically. ...

Abstract

Integrated photonic neural networks provide a promising platform for energy-efficient, high-throughput machine learning with extensive scientific and commercial applications. Photonic neural networks efficiently transform optically encoded inputs using Mach-Zehnder interferometer mesh networks interleaved with nonlinearities. We experimentally trained a three-layer, four-port silicon photonic neural network with programmable phase shifters and optical power monitoring to solve classification tasks using “in situ backpropagation,” a photonic analog of the most popular method to train conventional neural networks. We measured backpropagated gradients for phase-shifter voltages by interfering forward- and backward-propagating light and simulated in situ backpropagation for 64-port photonic neural networks trained on MNIST image recognition given errors. All experiments performed comparably to digital simulations (>94% test accuracy), and energy scaling analysis indicated a route to scalable machine learning."

Photonic Chips Curb AI Training’s Energy Appetite - IEEE Spectrum Stanford team achieves first-ever optical backpropagation milestone

Experimentally realized in situ backpropagation for deep learning in photonic neural networks (no public access)

World’s smallest LED could turn your phone camera into a high-res microscope

Amazing stuff! Smartphones are getting smarter? Lab on a chip smartphone!

"... researchers from the Singapore-MIT Alliance for Research and Technology (SMART) who’ve developed the world’s smallest silicon light-emitting diode (LED) – at less than a micrometer wide – with an intensity comparable to much larger silicon LEDs. ...
Previous on-chip emitters have been difficult to integrate into standard complementary metal-oxide-semiconductor (CMOS) platforms. ...
Here, the researchers placed their tiny silicon LED in a 55 nm CMOS node alongside the other photonic and electronic components – all on one chip. ...
To test how their LED might be used in a real-world situation, they placed it into a lensless holographic microscope. Lensless microscopes are smaller than regular microscopes and less expensive because they don’t require complex, precise lens systems. ...

the researchers used a neural networking algorithm to reconstruct objects viewed by the holographic microscope. ...

The researchers found that their holographic lens provided more accurate high-resolution images than a regular optical microscope. ..."

From the abstract:

"A nanoscale on-chip light source with high intensity is desired for various applications in integrated photonics systems. However, it is challenging to realize such an emitter using materials and fabrication processes compatible with the standard integrated circuit technology. In this letter, we report an electrically driven Si light-emitting diode with sub-wavelength emission area fabricated in an open-foundry microelectronics complementary metal-oxide-semiconductor platform. The light-emitting diode emission spectrum is centered around 1100 nm and the emission area is smaller than 0.14 μm2 (~∅400

 nm). This light-emitting diode has high spatial intensity of >50 mW/cm2 which is comparable with state-of-the-art Si-based emitters with much larger emission areas. Due to sub-wavelength confinement, the emission exhibits a high degree of spatial coherence, which is demonstrated by incorporating the light-emitting diode into a compact lensless in-line holographic microscope. This centimeter-scale, all-silicon microscope utilizes a single emitter to simultaneously illuminate ~9.5 million pixels of a complementary metal-oxide-semiconductor imager."

World’s smallest LED could turn your phone camera into a high-res microscope Researchers have created the world’s smallest silicon LED and holographic microscope that opens up a wide range of potential applications, including turning your smartphone camera into a portable, high-resolution microscope.

SMART researchers create world’s smallest LED and holographic microscope that enable conversion of existing mobile phone cameras into high-resolution microscopes

A sub-wavelength Si LED integrated in a CMOS platform (open access)

Fig. 1: Device structure and emission spectra.

Teaching photonic chips to do machine learning

Good news!

"... Photonic integrated circuits, or simply optical chips, have emerged as a possible solution to deliver higher computing performance, as measured by the number of operations performed per second per watt used, or TOPS/W. ..."

From the abstract:

"There has been growing interest in using photonic processors for performing neural network inference operations ... Here, we propose on-chip training of neural networks enabled by a CMOS-compatible silicon photonic architecture to harness the potential for massively parallel, efficient, and fast data operations. Our scheme employs the direct feedback alignment training algorithm, which trains neural networks using error feedback rather than error backpropagation, and can operate at speeds of trillions of multiply–accumulate (MAC) operations per second while consuming less than one picojoule per MAC operation. The photonic architecture exploits parallelized matrix–vector multiplications using arrays of microring resonators for processing multi-channel analog signals along single waveguide buses to calculate the gradient vector for each neural network layer in situ. We also experimentally demonstrate training deep neural networks with the MNIST dataset using on-chip MAC operation results. Our approach for efficient, ultra-fast neural network training showcases photonics as a promising platform for executing artificial intelligence applications."

Teaching photonic chips to 'learn' -- ScienceDaily (This article is an exact copy of the press release of the George Washington University below) A multi-institution research team has developed an optical chip that can train machine learning hardware.

Teaching photonic chips to learn A multi-institution research team has developed an optical chip that can train machine learning hardware. 

Silicon photonic architecture for training deep neural networks with direct feedback alignment (open access)

Fig. 1. Training photonic neural networks

New Optical Switch Could Lead to Ultrafast All-Optical Signal Processing

Good news! Doing business, education, and recreation at the speed of light!

"... Two things made the breakthrough possible: the material ... and the way in which they used it. First, they chose a crystalline material known as lithium niobate, a combination of niobium, lithium, and oxygen that does not occur in nature but has, over the past 50 years, proven essential to the field of optics. The material is inherently nonlinear: ...

more recently, advances in nanofabrication techniques have enabled ... to create lithium niobate-based integrated photonic devices that allow for the confinement of light in a tiny space. The smaller the space, the greater the intensity of light with the same amount of power. ...

confined the light temporally. Essentially, they decreased the duration of light pulses, and used a specific design that would keep the pulses short as they propagate through the device, which resulted in each pulse having higher peak power.

The combined effect of these two tactics—the spatiotemporal confinement of light—is to substantially enhance the strength of nonlinearity for a given pulse energy ...

The net result is the creation of a nonlinear splitter in which the light pulses are routed to two different outputs based on their energies, which enables switching to occur in less than 50 femtoseconds ... By comparison, state-of-the-art electronic switches take tens of picoseconds ..."

From the abstract:

"Optical nonlinear functions are crucial for various applications in integrated photonics, including all-optical information processing, photonic neural networks and on-chip ultrafast light sources. ... Here we effectively utilize the strong and instantaneous quadratic nonlinearity of lithium niobate nanowaveguides for the realization of cavity-free all-optical switching. By simultaneous engineering of the dispersion and quasi-phase matching, we design and demonstrate a nonlinear splitter that can achieve ultralow switching energies down to 80 fJ, featuring a fastest switching time of ~46 fs and a lowest energy–time product of 3.7 × 10−27 J s in integrated photonics. Our results can enable on-chip ultrafast and energy-efficient all-optical information processing, computing systems and light sources."

New Optical Switch Could Lead to Ultrafast All-Optical Signal Processing | www.caltech.edu Engineers at Caltech have developed a switch—one of the most fundamental components of computing—using optical, rather than electronic, components. The development could aid efforts to achieve ultrafast all-optical signal processing and computing.

Femtojoule femtosecond all-optical switching in lithium niobate nanophotonics (no public access)

Intel announces silicon photonics advancement towards optical I/O

Good news!

"Intel has demonstrated an eight-wavelength laser array on a silicon wafer. The research paves the way for the next generation of integrated silicon photonics products in the data center, such as switches with co-packaged optics and chiplets for optical interconnects. ...
More specifically, the laser uses a technology called dense wavelength division multiplexing (DWDM) in order to send different, closely spaced wavelengths over the same optical link. This technique thus increases the bandwidth while also reducing the physical size of the photonic chips. ..."

Intel announces silicon photonics advancement towards optical I/O | VentureBeat

A Chip That Can Process and Classify Nearly Two Billion Images per Second

Good news! Impressive! The future of photonic computation, a new computing paradigm! This could be a breakthrough, the first end-to-end system!

"... have removed the four main time-consuming culprits in the traditional computer chip: the conversion of optical to electrical signals, the need for converting the input data to binary format, a large memory module, and clock-based computations.
They have achieved this through direct processing of light received from the object of interest using an optical deep neural network implemented on a 9.3 square millimeter chip. ...
“When current computer chips process electrical signals they often run them through a Graphics Processing Unit, or GPU, which takes up space and energy,” ... “Our chip does not need to store the information, eliminating the need for a large memory unit. ...
“What’s really interesting about this technology is that it can do so much more than classify images,” ... “We already know how to convert many data types into the electrical domain – images, audio, speech, and many other data types. Now, we can convert different data types into the optical domain and have them processed almost instantaneously using this technology.” ...”

“And, by eliminating the memory unit that stores images, we are also increasing data privacy,” ..."

From the abstract:
"... In the optical domain, despite advances in photonic computation, the lack of scalable on-chip optical non-linearity and the loss of photonic devices limit the scalability of optical deep networks. Here we report an integrated end-to-end photonic deep neural network (PDNN) that performs sub-nanosecond image classification through direct processing of the optical waves impinging on the on-chip pixel array as they propagate through layers of neurons. In each neuron, linear computation is performed optically and the non-linear activation function is realized opto-electronically, allowing a classification time of under 570 ps, which is comparable with a single clock cycle of state-of-the-art digital platforms. A uniformly distributed supply light provides the same per-neuron optical output range, allowing scalability to large-scale PDNNs.  ..."

Penn Engineers Create Chip That Can Process and Classify Nearly Two Billion Images per Second

An on-chip photonic deep neural network for image classification (no public access)