Search

Recent advancements in artificial intelligence (AI) and deep neural networks (DNNs) have revolutionized numerous fields, enabling complex tasks by extracting intricate features from large datasets. However, the exponential growth in computational demands has outstripped the capabilities of traditional electrical hardware accelerators. Optical computing offers a promising alternative due to its inherent advantages of parallelism, high computational speed, and low power consumption. Yet, current photonic integrated circuits (PICs) designed for general matrix multiplication (GEMM) are constrained by large footprints, high costs of electro-optical (E-O) interfaces, and high control complexity, limiting their scalability. To overcome these challenges, we introduce a block-circulant photonic tensor core (CirPTC) for a structure-compressed optical neural network (StrC-ONN) architecture. By applying a structured compression strategy to weight matrices, StrC-ONN significantly reduces model parameters and hardware requirements while preserving the universal representability of networks and maintaining comparable expressivity. Additionally, we propose a hardware-aware training framework to compensate for on-chip nonidealities to improve model robustness and accuracy. We experimentally demonstrate image processing and classification tasks, achieving up to a 74.91% reduction in trainable parameters while maintaining competitive accuracies. Performance analysis expects a computational density of 5.84 tera operations per second (TOPS) per mm^2 and a power efficiency of 47.94 TOPS/W, marking a 6.87-times improvement achieved through the hardware-software co-design approach. By reducing both hardware requirements and control complexity across multiple dimensions, this work explores a new pathway to push the limits of optical computing in the pursuit of high efficiency and scalability.

Electromagnetic field simulation is central to designing, optimizing, and validating photonic devices and circuits. However, costly computation associated with numerical simulation poses a significant bottleneck, hindering scalability and turnaround time in the photonic circuit design process. Neural operators offer a promising alternative, but existing SOTA approaches, NeurOLight, struggle with predicting high-fidelity fields for real-world complicated photonic devices, with the best reported 0.38 normalized mean absolute error in NeurOLight. The inter-plays of highly complex light-matter interaction, e.g., scattering and resonance, sensitivity to local structure details, non-uniform learning complexity for full-domain simulation, and rich frequency information, contribute to the failure of existing neural PDE solvers. In this work, we boost the prediction fidelity to an unprecedented level for simulating complex photonic devices with a novel operator design driven by the above challenges. We propose a novel cross-axis factorized PACE operator with a strong long-distance modeling capacity to connect the full-domain complex field pattern with local device structures. Inspired by human learning, we further divide and conquer the simulation task for extremely hard cases into two progressively easy tasks, with a first-stage model learning an initial solution refined by a second model. On various complicated photonic device benchmarks, we demonstrate one sole PACE model is capable of achieving 73% lower error with 50% fewer parameters compared with various recent ML for PDE solvers. The two-stage setup further advances high-fidelity simulation for even more intricate cases. In terms of runtime, PACE demonstrates 154-577x and 11.8-12x simulation speedup over numerical solver using scipy or highly-optimized pardiso solver, respectively. We open sourced the code and dataset., Comment: Accepeted by Neurips 2024, 21 pages

We analyze the mixing time of a popular shuffling machine known as the shelf shuffler. It is a modified version of a $2m$-handed riffle shuffle ($m=10$ in casinos) in which a deck of $n$ cards is split multinomially into $2m$ piles, the even-numbered piles are reversed, and then cards are dropped from piles proportionally to their sizes. We prove that $\frac{5}{4} \log_{2m} n$ shuffles are necessary and sufficient to mix in total variation, and a cutoff occurs with constant window size. We also determine the cutoff profile in terms of the total variation distance between two shifted normal random variables., Comment: 1 figure

The extension of photonic technologies such as lidar and free-space optical communications from the traditional visible and near-infrared wavelengths to longer wavelengths can improve performance in adverse environments such as haze, fog, smoke, or strong solar background. Non-mechanical beam steerers will be a critical component of the low size, weight, and power modules needed for the portable or unmanned systems deployed in these environments. In this work, we demonstrate the first 2D optical phased array for non-mechanical beam steering in the mid-infrared spectral region. We combine a total internal-reflection mirror emitter with a non-redundant array of 30 elements to carry out 2D beam steering at a single wavelength of 4.6 $\mu$m. The experiment yielded 600 resolvable far-field points, with 2400 points over a $28^\circ \times 28^\circ$ field of view calculated theoretically. Moreover, the device was fabricated in a passive InGaAs/InP platform, contributing another advance in the ongoing development of quantum cascade laser-based photonic integration., Comment: Main document: 16 pages, 11 figures; Supplement: 3 pages, 2 figures

As photonic technologies continue to grow in multidimensional aspects, integrated photonics holds a unique position and continuously presents enormous possibilities to research communities. Applications span across data centers, environmental monitoring, medical diagnosis, and highly compact communication components, with further possibilities growing endlessly. Here, we provide a review of state of the art integrated photonic sensors operating in near and mid infrared wavelength regions on various material platforms. Among different materials, architectures, and technologies leading the way for on chip sensors, we discuss optical sensing principles commonly applied to biochemical and gas sensing. Our focus is particularly on passive and active optical waveguides, including dispersion engineered metamaterial based structures an essential approach for enhancing the interaction between light and analytes in chip scale sensors. We harness a diverse array of cutting edge sensing technologies, heralding a revolutionary on chip sensing paradigm. Our arsenal includes refractive index based sensing, plasmonic, and spectroscopy, forging an unparalleled foundation for innovation and precision. Furthermore, we include a brief discussion of recent trends and computational concepts incorporating Artificial Intelligence & Machine Learning (AI/ML) and deep learning approaches over the past few years to improve the qualitative and quantitative analysis of sensor measurements.

In recent decades, the demand for computational power has surged, particularly with the rapid expansion of artificial intelligence (AI). As we navigate the post-Moore's law era, the limitations of traditional electrical digital computing, including process bottlenecks and power consumption issues, are propelling the search for alternative computing paradigms. Among various emerging technologies, integrated photonics stands out as a promising solution for next-generation high-performance computing, thanks to the inherent advantages of light, such as low latency, high bandwidth, and unique multiplexing techniques. Furthermore, the progress in photonic integrated circuits (PICs), which are equipped with abundant photoelectronic components, positions photonic-electronic integrated circuits as a viable solution for high-performance computing and hardware AI accelerators. In this review, we survey recent advancements in both PIC-based digital and analog computing for AI, exploring the principal benefits and obstacles of implementation. Additionally, we propose a comprehensive analysis of photonic AI from the perspectives of hardware implementation, accelerator architecture, and software-hardware co-design. In the end, acknowledging the existing challenges, we underscore potential strategies for overcoming these issues and offer insights into the future drivers for optical computing.

The optical neural network (ONN) is a promising hardware platform for next-generation neuromorphic computing due to its high parallelism, low latency, and low energy consumption. However, previous integrated photonic tensor cores (PTCs) consume numerous single-operand optical modulators for signal and weight encoding, leading to large area costs and high propagation loss to implement large tensor operations. This work proposes a scalable and efficient optical dot-product engine based on customized multi-operand photonic devices, namely multi-operand optical neurons (MOON). We experimentally demonstrate the utility of a MOON using a multi-operand-Mach-Zehnder-interferometer (MOMZI) in image recognition tasks. Specifically, our MOMZI-based ONN achieves a measured accuracy of 85.89% in the street view house number (SVHN) recognition dataset with 4-bit voltage control precision. Furthermore, our performance analysis reveals that a 128x128 MOMZI-based PTCs outperform their counterparts based on single-operand MZIs by one to two order-of-magnitudes in propagation loss, optical delay, and total device footprint, with comparable matrix expressivity., Comment: 19 pages, 10 figures

The wide adoption and significant computing resource of attention-based transformers, e.g., Vision Transformers and large language models (LLM), have driven the demand for efficient hardware accelerators. There is a growing interest in exploring photonics as an alternative technology to digital electronics due to its high energy efficiency and ultra-fast processing speed. Photonic accelerators have shown promising results for CNNs, which mainly rely on weight-static linear operations. However, they encounter issues when efficiently supporting Transformer architectures, questioning the applicability of photonics to advanced ML tasks. The primary hurdle lies in their inefficiency in handling unique workloads in Transformers, i.e., dynamic and full-range tensor multiplication. In this work, we propose Lightening-Transformer, the first light-empowered, high-performance, and energy-efficient photonic Transformer accelerator. To overcome prior designs' fundamental limitations, we introduce a novel dynamically-operated photonic tensor core, DPTC, a crossbar array of interference-based optical vector dot-product engines supporting highly parallel, dynamic, and full-range matrix multiplication. Furthermore, we design a dedicated accelerator that integrates our novel photonic computing cores with photonic interconnects for inter-core data broadcast, fully unleashing the power of optics. Comprehensive evaluations show that ours achieves >2.6x energy and >12x latency reductions compared to prior photonic accelerators and delivers the lowest energy cost and 2 to 3 orders of magnitude lower energy-delay product compared to electronic Transformer accelerators, all while maintaining digital-comparable accuracy. Our work highlights the immense potential of photonics for advanced ML workloads, such as Transformer-backboned LLM. Our work is available at https://github.com/zhuhanqing/Lightening-Transformer., Comment: Published as a conference paper in HPCA 2024. Recieved the Reproducibility Badges at IEEE. Our implementation is available at https://github.com/zhuhanqing/Lightening-Transformer

Photonic computing shows promise for transformative advancements in machine learning (ML) acceleration, offering ultra-fast speed, massive parallelism, and high energy efficiency. However, current photonic tensor core (PTC) designs based on standard optical components hinder scalability and compute density due to their large spatial footprint. To address this, we propose an ultra-compact PTC using customized programmable multi-operand multimode interference (MOMMI) devices, named M3ICRO. The programmable MOMMI leverages the intrinsic light propagation principle, providing a single-device programmable matrix unit beyond the conventional computing paradigm of one multiply-accumulate (MAC) operation per device. To overcome the optimization difficulty of customized devices that often requires time-consuming simulation, we apply ML for optics to predict the device behavior and enable a differentiable optimization flow. We thoroughly investigate the reconfigurability and matrix expressivity of our customized PTC, and introduce a novel block unfolding method to fully exploit the computing capabilities of a complex-valued PTC for near-universal real-valued linear transformations. Extensive evaluations demonstrate that M3ICRO achieves a 3.4-9.6x smaller footprint, 1.6-4.4x higher speed, 10.6-42x higher compute density, 3.7-12x higher system throughput, and superior noise robustness compared to state-of-the-art coherent PTC designs, while maintaining close-to-digital task accuracy across various ML benchmarks. Our code is open-sourced at https://github.com/JeremieMelo/M3ICRO-MOMMI., Comment: 12 pages. Accepted to APL Machine Learning 2023

In the context of continued spread of coronavirus disease 2019 (COVID-19) caused by SARS-CoV-2 and the emergence of new variants, the demand for rapid, accurate, and frequent detection is increasing. Besides, the new predominant strain, Omicron variant, manifests more similar clinical features to those of other common respiratory infections. The concurrent detection of multiple potential pathogens helps distinguish SARS-CoV-2 infection from other diseases with overlapping symptoms, which is significant for patients to receive tailored treatment and containing the outbreak. Here, we report a lab-on-a-chip biosensing platform for SARS-CoV-2 detection based on subwavelength grating micro-ring resonator. The sensing surface is functionalized by specific antibody against SARS-CoV-2 spike protein, which could produce redshifts of resonant peaks by antigen-antibody combination, thus achieving quantitative detection. Additionally, the sensor chip is integrated with a microfluidic chip with an anti-backflow Y-shaped structure that enables the concurrent detection of two analytes. In this study, we realized the detection and differentiation of COVID-19 and influenza A H1N1. Experimental results show that the limit of detection of our device reaches 100 fg/mL (1.31 fM) within 15 min detecting time, and cross-reactivity tests manifest the specificity of the optical diagnostic assay. Further, the integrated packaging and streamlined workflow facilitate its use for clinical applications. Thus, the biosensing platform offers a promising solution to achieve ultrasensitive, selective, multiplexed, and quantitative point-of-care detection of COVID-19.

Optical computing is an emerging technology for next-generation efficient artificial intelligence (AI) due to its ultra-high speed and efficiency. Electromagnetic field simulation is critical to the design, optimization, and validation of photonic devices and circuits. However, costly numerical simulation significantly hinders the scalability and turn-around time in the photonic circuit design loop. Recently, physics-informed neural networks have been proposed to predict the optical field solution of a single instance of a partial differential equation (PDE) with predefined parameters. Their complicated PDE formulation and lack of efficient parametrization mechanisms limit their flexibility and generalization in practical simulation scenarios. In this work, for the first time, a physics-agnostic neural operator-based framework, dubbed NeurOLight, is proposed to learn a family of frequency-domain Maxwell PDEs for ultra-fast parametric photonic device simulation. We balance the efficiency and generalization of NeurOLight via several novel techniques. Specifically, we discretize different devices into a unified domain, represent parametric PDEs with a compact wave prior, and encode the incident light via masked source modeling. We design our model with parameter-efficient cross-shaped NeurOLight blocks and adopt superposition-based augmentation for data-efficient learning. With these synergistic approaches, NeurOLight generalizes to a large space of unseen simulation settings, demonstrates 2-orders-of-magnitude faster simulation speed than numerical solvers, and outperforms prior neural network models by ~54% lower prediction error with ~44% fewer parameters. Our code is available at https://github.com/JeremieMelo/NeurOLight., Comment: 13 pages. Accepted to NeurIPS 2022

Analog computing has been recognized as a promising low-power alternative to digital counterparts for neural network acceleration. However, conventional analog computing is mainly in a mixed-signal manner. Tedious analog/digital (A/D) conversion cost significantly limits the overall system's energy efficiency. In this work, we devise an efficient analog activation unit with magnetic tunnel junction (MTJ)-based analog content-addressable memory (MACAM), simultaneously realizing nonlinear activation and A/D conversion in a fused fashion. To compensate for the nascent and therefore currently limited representation capability of MACAM, we propose to mix our analog activation unit with digital activation dataflow. A fully differential framework, SuperMixer, is developed to search for an optimized activation workload assignment, adaptive to various activation energy constraints. The effectiveness of our proposed methods is evaluated on a silicon photonic accelerator. Compared to standard activation implementation, our mixed activation system with the searched assignment can achieve competitive accuracy with $>$60% energy saving on A/D conversion and activation., Comment: Accepted by ICCAD 2022

A micro-ring-resonator (MRR) optical biosensor based on the evanescent field sensing mechanism has been extensively studied due to its high sensitivity and compact device size. However, a suitable on-chip integrated spectrometer device has to be demonstrated for the lab-on-a-chip applications, which can read the resonance wavelength shift from MRR biosensors based on minuscule changes in refractive index. In this paper, we demonstrated the design and experimental results of the near-infrared lab-on-a-chip optical biosensor platform that monolithically integrates the MRR and the on-chip spectrometer on the silicon-on-insulator (SOI) wafer, which can eliminate the external optical spectrum analyzer for scanning the wavelength spectrum. The symmetric add-drop MRR biosensor is designed to have a free spectral range (FSR) of ~19 nm, and a bulk sensitivity of ~73 nm/RIU; then the drop-port output resonance peaks are reconstructed from the integrated spatial-heterodyne Fourier transform spectrometer (SHFTS) with the spectral resolution of ~3.1 nm and bandwidth of ~50 nm, which results in the limit of detection of 0.042 RIU. The MRR output spectrum with air- and water-claddings are measured and reconstructed from the MRR-SHFTS integrated device experimentally to validate the wavelength shifting measurement., Comment: 23 pages, 9 figures including supplementary

Optical I/O plays a crucial role in the lifespan of lab-on-a-chip systems, from preliminary testing to operation in the target environment. However, due to the precise alignments required, efficient and reliable fiber-to-chip connections remain challenging, yielding inconsistent test results and unstable packaged performance. To overcome this issue, for use in single mode on-chip systems, we propose the incorporation of area-enlarged grating couplers working in conjunction with multimode fibers. This combination enables simpler, faster, and more reliable connections than the traditional small area grating coupler with single-mode fiber. In this work, we experimentally demonstrate a 3dB in-plane (X, Y) spatial tolerance of (10.2 {\mu}m, 17.3 {\mu}m) for the large area configuration, being at least (2.49, 3.33) times that of the small area one, and agreeing well with theoretical calculations. The simple concept is readily applicable to a range of photonic systems where cheaper more robust optical I/O is desired., Comment: 11 pages, 5 figures

Photonic tensor cores (PTCs) are essential building blocks for optical artificial intelligence (AI) accelerators based on programmable photonic integrated circuits. PTCs can achieve ultra-fast and efficient tensor operations for neural network (NN) acceleration. Current PTC designs are either manually constructed or based on matrix decomposition theory, which lacks the adaptability to meet various hardware constraints and device specifications. To our best knowledge, automatic PTC design methodology is still unexplored. It will be promising to move beyond the manual design paradigm and "nurture" photonic neurocomputing with AI and design automation. Therefore, in this work, for the first time, we propose a fully differentiable framework, dubbed ADEPT, that can efficiently search PTC designs adaptive to various circuit footprint constraints and foundry PDKs. Extensive experiments show superior flexibility and effectiveness of the proposed ADEPT framework to explore a large PTC design space. On various NN models and benchmarks, our searched PTC topology outperforms prior manually-designed structures with competitive matrix representability, 2-30x higher footprint compactness, and better noise robustness, demonstrating a new paradigm in photonic neural chip design. The code of ADEPT is available at https://github.com/JeremieMelo/ADEPT using the https://github.com/JeremieMelo/pytorch-onn (TorchONN) library., Comment: Accepted to ACM/IEEE Design Automation Conference (DAC), 2022

With the recent advances in optical phase change material (PCM), photonic in-memory neurocomputing has demonstrated its superiority in optical neural network (ONN) designs with near-zero static power consumption, time-of-light latency, and compact footprint. However, photonic tensor cores require massive hardware reuse to implement large matrix multiplication due to the limited single-core scale. The resultant large number of PCM writes leads to serious dynamic power and overwhelms the fragile PCM with limited write endurance. In this work, we propose a synergistic optimization framework, ELight, to minimize the overall write efforts for efficient and reliable optical in-memory neurocomputing. We first propose write-aware training to encourage the similarity among weight blocks, and combine it with a post-training optimization method to reduce programming efforts by eliminating redundant writes. Experiments show that ELight can achieve over 20X reduction in the total number of writes and dynamic power with comparable accuracy. With our ELight, photonic in-memory neurocomputing will step forward towards viable applications in machine learning with preserved accuracy, order-of-magnitude longer lifetime, and lower programming energy., Comment: 7 pages, 8 figures, accepted by ASPDAC 2022

The on-chip broadband optical spectrometers which cover the entire tissue transparency window ({\lambda}=650-1050 nm) with high resolution are highly demanded for the miniaturized bio-sensing and bio-imaging applications. Here, we propose a novel type of spatial heterodyne Fourier transform spectrometer (SHFTS) integrated with a sub-wavelength grating coupler (SWGC) for the dual-polarization bandpass sampling on the Si3N4 platform. Through tuning the coupling angles with different polarization, we experimentally demonstrated the unprecedented broadband spectrum retrieval results with the overall bandwidth coverage of 400 nm, bridging the wavelengths from 650 nm to 1050 nm, with the resolution of 2-5 nm. By applying the bandpass sampling theorem, we circumvented the intrinsic trade-off limitation between the bandwidth and resolution of SHFTS without having an outrageous number of Mach-Zehnder interferometer (MZI) arrays or adding additional active components. The bandpass sampling SHFTS is designed to have linearly unbalanced 32 MZIs with the maximum optical path length difference of 93 {\mu}m within an overall footprint size of 4.7 mm x 0.65 mm, and the coupling angles of SWGC are varied from 0{\deg} to 32{\deg} to cover the entire tissue transparency window., Comment: 48 Pages, 6 figures, 14 supportive figures

The optical neural network (ONN) is a promising hardware platform for next-generation neurocomputing due to its high parallelism, low latency, and low energy consumption. Previous ONN architectures are mainly designed for general matrix multiplication (GEMM), leading to unnecessarily large area cost and high control complexity. Here, we move beyond classical GEMM-based ONNs and propose an optical subspace neural network (OSNN) architecture, which trades the universality of weight representation for lower optical component usage, area cost, and energy consumption. We devise a butterfly-style photonic-electronic neural chip to implement our OSNN with up to 7x fewer trainable optical components compared to GEMM-based ONNs. Additionally, a hardware-aware training framework is provided to minimize the required device programming precision, lessen the chip area, and boost the noise robustness. We experimentally demonstrate the utility of our neural chip in practical image recognition tasks, showing that a measured accuracy of 94.16% can be achieved in hand-written digit recognition tasks with 3-bit weight programming precision., Comment: 17 pages,5 figures

Silicon-photonics-based optical neural network (ONN) is a promising hardware platform that could represent a paradigm shift in efficient AI with its CMOS-compatibility, flexibility, ultra-low execution latency, and high energy efficiency. In-situ training on the online programmable photonic chips is appealing but still encounters challenging issues in on-chip implementability, scalability, and efficiency. In this work, we propose a closed-loop ONN on-chip learning framework L2ight to enable scalable ONN mapping and efficient in-situ learning. L2ight adopts a three-stage learning flow that first calibrates the complicated photonic circuit states under challenging physical constraints, then performs photonic core mapping via combined analytical solving and zeroth-order optimization. A subspace learning procedure with multi-level sparsity is integrated into L2ight to enable in-situ gradient evaluation and fast adaptation, unleashing the power of optics for real on-chip intelligence. Extensive experiments demonstrate our proposed L2ight outperforms prior ONN training protocols with 3-order-of-magnitude higher scalability and over 30X better efficiency, when benchmarked on various models and learning tasks. This synergistic framework is the first scalable on-chip learning solution that pushes this emerging field from intractable to scalable and further to efficient for next-generation self-learnable photonic neural chips. From a co-design perspective, L2ight also provides essential insights for hardware-restricted unitary subspace optimization and efficient sparse training. We open-source our framework at https://github.com/JeremieMelo/L2ight., Comment: 10 pages. Accepted to NeurIPS 2021

Deep neural networks (DNN) have shown superior performance in a variety of tasks. As they rapidly evolve, their escalating computation and memory demands make it challenging to deploy them on resource-constrained edge devices. Though extensive efficient accelerator designs, from traditional electronics to emerging photonics, have been successfully demonstrated, they are still bottlenecked by expensive memory accesses due to tremendous gaps between the bandwidth/power/latency of electrical memory and computing cores. Previous solutions fail to fully-leverage the ultra-fast computational speed of emerging DNN accelerators to break through the critical memory bound. In this work, we propose a general and unified framework to trade expensive memory transactions with ultra-fast on-chip computations, directly translating to performance improvement. We are the first to jointly explore the intrinsic correlations and bit-level redundancy within DNN kernels and propose a multi-level in situ generation mechanism with mixed-precision bases to achieve on-the-fly recovery of high-resolution parameters with minimum hardware overhead. Extensive experiments demonstrate that our proposed joint method can boost the memory efficiency by 10-20x with comparable accuracy over four state-of-the-art designs, when benchmarked on ResNet-18/DenseNet-121/MobileNetV2/V3 with various tasks., Comment: Accepted by International Conference on Computer Vision (ICCV) 2021

We develop a variational Bayesian (VB) approach for estimating large-scale dynamic network models in the network autoregression framework. The VB approach allows for the automatic identification of the dynamic structure of such a model and obtains a direct approximation of the posterior density. Compared to Markov Chain Monte Carlo (MCMC) based sampling approaches, the VB approach achieves enhanced computational efficiency without sacrificing estimation accuracy. In the simulation study conducted here, the proposed VB approach detects various types of proper active structures for dynamic network models. Compared to the alternative approach, the proposed method achieves similar or better accuracy, and its computational time is halved. In a real data analysis scenario of day-ahead natural gas flow prediction in the German gas transmission network with 51 nodes between October 2013 and September 2015, the VB approach delivers promising forecasting accuracy along with clearly detected structures in terms of dynamic dependence., Comment: 40 pages, 38 figures

Optical neural networks (ONNs) have demonstrated record-breaking potential in high-performance neuromorphic computing due to their ultra-high execution speed and low energy consumption. However, current learning protocols fail to provide scalable and efficient solutions to photonic circuit optimization in practical applications. In this work, we propose a novel on-chip learning framework to release the full potential of ONNs for power-efficient in situ training. Instead of deploying implementation-costly back-propagation, we directly optimize the device configurations with computation budgets and power constraints. We are the first to model the ONN on-chip learning as a resource-constrained stochastic noisy zeroth-order optimization problem, and propose a novel mixed-training strategy with two-level sparsity and power-aware dynamic pruning to offer a scalable on-chip training solution in practical ONN deployment. Compared with previous methods, we are the first to optimize over 2,500 optical components on chip. We can achieve much better optimization stability, 3.7x-7.6x higher efficiency, and save >90% power under practical device variations and thermal crosstalk., Comment: 7 pages content, 2 pages of reference, 6 figures, 4 tables, accepted to Association for the Advancement of Artificial Intelligence (AAAI) 2021

The vertical-cavity surface-emitting lasers (VCSELs) have emerged as a vital approach for realizing energy efficient, high speed optical interconnects in the data center and supercomputers. As of today, VCSEL is the most suitable for mass production in terms of cost-effectiveness and reliability. However, there are still key challenges for higher speed modulation above 40 GHz. Here, a hexagonal transverse coupled cavity VCSEL adiabatically coupled through the center cavity is proposed. A 3-dB roll-off modulation bandwidth of 45 GHz is demonstrated, which is five times greater than a conventional VCSEL fabricated on the same epi-wafer structure. While a parity time (PT) symmetry approaches add loss to engineer the topological state of the laser system, here, a radical paradigm shift with gain introduces symmetry breaking. This idea, then enables a single mode operation with a side-mode suppression-ratio (SMSR) of > 30 decibels and signal-to-noise ratio (SNR) of > 45 decibels. The energy distribution inside the coupled cavity system is also redistributed to provide a coherent gain in a spatially separated system. Consequently, throughput power is three times higher than that of the conventional VCSEL., Comment: 14 pages, 4 figures

The sudden rise of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic early 2020 throughout the world has called into drastic action measures to do instant detection and reduce the spread rate. The common diagnostics testing methods has been only partially effective in satisfying the booming demand for fast detection methods to contain the further spread. However, the point-of-risk accurate diagnosis of this new emerging viral infection is paramount as simultaneous normal working operation and dealing with symptoms of SARS-CoV-2 can become the norm for years to come. Sensitive cost-effective biosensor with mass production capability is crucial throughout the world until a universal vaccination become available. Optical label-free biosensors can provide a non-invasive, extremely sensitive rapid detection technique up to ~1 fM concentration along with few minutes sensing. These biosensors can be manufactured on a mass-scale (billions) to detect the COVID-19 viral load in nasal, saliva, urinal, and serological samples even if the infected person is asymptotic. Methods investigated here are the most advanced available platforms for biosensing optical devices resulted from the integration of state-of-the-art designs and materials. These approaches are including but not limited to integrated optical devices, plasmonic resonance and also emerging nanomaterial biosensors. The lab-on-a-chip platforms examined here are suitable not only for SARS-CoV-2 spike protein detection but also other contagious virions such as influenza, and middle east respiratory syndrome (MERS)., Comment: 52 pages, 19 figures

Densely integrated active photonics is key for next generation on-chip networks for addressing both footprint and energy budget concerns. However, the weak light-matter interaction in traditional active Silicon optoelectronics mandates rather sizable device lengths. The ideal active material choice should avail high index modulation while being easily integrated into Silicon photonics platforms. Indium tin oxide (ITO) offers such functionalities and has shown promising modulation capacity recently. Interestingly, the nanometer-thin unity-strong index modulation of ITO synergistically combines the high group-index in hybrid plasmonic with nanoscale optical modes. Following this design paradigm, here, we demonstrate a spectrally broadband, GHz-fast Mach-Zehnder interferometric modulator, exhibiting a high efficiency signified by a miniscule VpL of 95 Vum, deploying an one-micrometer compact electrostatically tunable plasmonic phase-shifter, based on heterogeneously integrated ITO thin films into silicon photonics. Furthermore we show, that this device paradigm enables spectrally broadband operation across the entire telecommunication near infrared C-band. Such sub-wavelength short efficient and fast modulators monolithically integrated into Silicon platform open up new possibilities for high-density photonic circuitry, which is critical for high interconnect density of photonic neural networks or applications in GHz-fast optical phased-arrays, for example.

We introduce a comprehensive screening platform for the COVID-19 (a.k.a., SARS-CoV-2) pneumonia. The proposed AI-based system works on chest x-ray (CXR) images to predict whether a patient is infected with the COVID-19 disease. Although the recent international joint effort on making the availability of all sorts of open data, the public collection of CXR images is still relatively small for reliably training a deep neural network (DNN) to carry out COVID-19 prediction. To better address such inefficiency, we design a cascaded learning strategy to improve both the sensitivity and the specificity of the resulting DNN classification model. Our approach leverages a large CXR image dataset of non-COVID-19 pneumonia to generalize the original well-trained classification model via a cascaded learning scheme. The resulting screening system is shown to achieve good classification performance on the expanded dataset, including those newly added COVID-19 CXR images., Comment: 14 pages, 6 figures

Global optimization of expensive functions has important applications in physical and computer experiments. It is a challenging problem to develop efficient optimization scheme, because each function evaluation can be costly and the derivative information of the function is often not available. We propose a novel global optimization framework using adaptive Radial Basis Functions (RBF) based surrogate model via uncertainty quantification. The framework consists of two iteration steps. It first employs an RBF-based Bayesian surrogate model to approximate the true function, where the parameters of the RBFs can be adaptively estimated and updated each time a new point is explored. Then it utilizes a model-guided selection criterion to identify a new point from a candidate set for function evaluation. The selection criterion adopted here is a sample version of the expected improvement (EI) criterion. We conduct simulation studies with standard test functions, which show that the proposed method has some advantages, especially when the true surface is not very smooth. In addition, we also propose modified approaches to improve the search performance for identifying global optimal points and to deal with the higher dimension scenarios., Comment: 35 pages, 5 figures, 5 tables

Here, we demonstrate a spectrally broadband, GHz-fast Mach-Zehnder interferometeric modulator, exhibiting a miniscule VpL of 95 V-um, deploying a sub-wavelength short electrostatically tunable plasmonic phase-shifter, based on heterogeneously integrated ITO thin films into silicon photonics.

Real-time tweets can provide useful information on evolving events and situations. Geotagged tweets are especially useful, as they indicate the location of origin and provide geographic context. However, only a small portion of tweets are geotagged, limiting their use for situational awareness. In this paper, we adapt, improve, and evaluate a state-of-the-art deep learning model for city-level geolocation prediction, and integrate it with a visual analytics system tailored for real-time situational awareness. We provide computational evaluations to demonstrate the superiority and utility of our geolocation prediction model within an interactive system., Comment: 4 pages, 2 tables, 1 figure, SIGSPATIAL GeoAI Workshop

Electro-optic signal modulation provides a key functionality in modern technology and information networks. Photonic integration has enabled not only miniaturizing photonic components, but also provided performance improvements due to co-design addressing both electrical and optical device rules. However, the millimeter-to-centimeter large footprint of many foundry-ready photonic electro-optic modulators significantly limits on-chip scaling density. To address these limitations, here we experimentally demonstrate a coupling-enhanced electro-absorption modulator by heterogeneously integrating a novel dual-gated indium-tin-oxide (ITO) phase-shifting tunable absorber placed at a silicon directional coupler region. Our experimental modulator shows a 2 dB extinction ratio for a just 4 um short device at 4 volt bias. Since no material nor optical resonances are deployed, this device shows spectrally broadband operation as demonstrated here across the entire C-band. In conclusion we demonstrate a modulator utilizing strong index-change from both real and imaginary part of active material enabling compact and high-performing modulators using semiconductor foundry-near materials.

In a laser system non-Hermitian methods such as Parity-Time (PT) Symmetry and Supersymmetry (SUSY) have shown and demonstrated the ability to suppress unwanted lasing modes and, thus, achieved single mode lasing operation through the addition of lossy passive elements. While these approaches enable laser engineering versatility, they rely on the drawback of adding optical losses to a system tasked to produce single mode gain. Unlike PT and SUSY lasers, here we show an extra loss-free non-Hermitian laser engineering approach to realize single mode lasing operation for the first time. By selectively enhancing the fundamental modes quality factor, we obtain single mode operation with higher output power per cavity since all cavities in this system contribute to the laser output, in contrast to other non-Hermitian approaches. Furthermore, we show that this approach interestingly allows reducing the number of to-be-designed cavities in super-partner array as compared with, for example, the SUSY approach, thus leading to reduced design complexity upon coupled cavity scale up of laser arrays. In summary, the ability to engineer coupled laser systems where each laser cavity contributes to coherent light amplification opens up a new degree of laser-design freedom leading to increased device performance and simultaneous reduced design and fabrication complexity.

We investigate 1D aperiodic multilayer microstructures in order to achieve near total absorption in preselected wavelengths in a graphene monolayer. Our structures are designed by a genetic optimization algorithm coupled to a transfer matrix code. Coupled mode theory (CMT) analysis, in accordance with transfer matrix method (TMM) results, indicates the existence of a critical coupling in a graphene monolayer for perfect absorptions. Our findings show that the near-total-absorption peaks are highly tunable and can be controlled simultaneously or independently in wide range of wavelengths in the near-infrared and visible. Our proposed approach is metal free and does not require surface texturing or patterning, and can be applied for other two dimensional (2D) materials., Comment: 12 pages, 5 figures

We study a logistic model-based active learning procedure for binary classification problems, in which we adopt a batch subject selection strategy with a modified sequential experimental design method. Moreover, accompanying the proposed subject selection scheme, we simultaneously conduct a greedy variable selection procedure such that we can update the classification model with all labeled training subjects. The proposed algorithm repeatedly performs both subject and variable selection steps until a prefixed stopping criterion is reached. Our numerical results show that the proposed procedure has competitive performance, with smaller training size and a more compact model, comparing with that of the classifier trained with all variables and a full data set. We also apply the proposed procedure to a well-known wave data set (Breiman et al., 1984) to confirm the performance of our method.

In atomic multi-level systems, adiabatic elimination is a method used to minimize complicity of the system by eliminating irrelevant and strongly coupled levels by detuning them from one-another. Such a three-level system, for instance, can be mapped onto physical in form of a three-waveguide system. Actively detuning the coupling strength between the respective waveguide modes allows modulating light propagating through the device, as proposed here. The outer waveguides act as an effective two- photonic-mode system similar to ground- and excited states of a three-level atomic system, whilst the center waveguide is partially plasmonic. In adiabatic elimination regime, the amplitude of the middle waveguide oscillates much faster in comparison to the outer waveguides leading to a vanishing field build up. As a result, the middle waveguide becomes a dark state and hence a low insertion-loss of 8 decibel is expected to keep when achieving the modulation depth as high as 70 decibel despite the involvement of a plasmonic waveguide in the design presented here. The modulation mechanism relies on switching this waveguide system from a critical coupling regime to adiabatic elimination condition via electrostatically tuning the free-carrier concentration and hence the optical index of a thin ITO layer residing in the plasmonic center waveguide. This alters the effective coupling length and the phase mismatching condition thus modulation in each of outer waveguides. Our results show a modulator energy efficiency as low as 40 atto-joule per bit and an extinction ratio of 50 decibel. Given the minuscule footprint of the modulator, the resulting lumped-element limited RC delay is expected to exceed 200 giga hertz. This type of modulator paves the way for next-generation both energy-and speed conscience optical short-reach interconnects., Comment: 11 Pages,5 figures

In this work, we consider a manufactory process which can be described by a multiple-instance logistic regression model. In order to compute the maximum likelihood estimation of the unknown coefficient, an expectation-maximization algorithm is proposed, and the proposed modeling approach can be extended to identify the important covariates by adding the coefficient penalty term into the likelihood function. In addition to essential technical details, we demonstrate the usefulness of the proposed method by simulations and real examples.