Explore the vast range of titles available.

Matrix computation, as a fundamental building block of information processing in science and technology, contributes most of the computational overheads in modern signal processing and artificial intelligence algorithms. Photonic accelerators are designed to accelerate specific categories of computing in the optical domain, especially matrix multiplication, to address the growing demand for computing resources and capacity. Photonic matrix multiplication has much potential to expand the domain of telecommunication, and artificial intelligence benefiting from its superior performance. Recent research in photonic matrix multiplication has flourished and may provide opportunities to develop applications that are unachievable at present by conventional electronic processors. In this review, we first introduce the methods of photonic matrix multiplication, mainly including the plane light conversion method, Mach–Zehnder interferometer method and wavelength division multiplexing method. We also summarize the developmental milestones of photonic matrix multiplication and the related applications. Then, we review their detailed advances in applications to optical signal processing and artificial neural networks in recent years. Finally, we comment on the challenges and perspectives of photonic matrix multiplication and photonic acceleration.This review summarizes the advances of photonic accelerators from the viewpoint of photonic matrix multiplication, providing a guidance for all-optical or optoelectronic-hybrid AI hardware chip system.

Multimodal deep learning plays a pivotal role in supporting the processing and learning of diverse data types within the realm of artificial intelligence generated content (AIGC). However, most photonic neuromorphic processors for deep learning can only handle a single data modality (either vision or audio) due to the lack of abundant parameter training in optical domain. Here, we propose and demonstrate a trainable diffractive optical neural network (TDONN) chip based on on-chip diffractive optics with massive tunable elements to address these constraints. The TDONN chip includes one input layer, five hidden layers, and one output layer, and only one forward propagation is required to obtain the inference results without frequent optical-electrical conversion. The customized stochastic gradient descent algorithm and the drop-out mechanism are developed for photonic neurons to realize in situ training and fast convergence in the optical domain. The TDONN chip achieves a potential throughput of 217.6 tera-operations per second (TOPS) with high computing density (447.7 TOPS/mm 2 ), high system-level energy efficiency (7.28 TOPS/W), and low optical latency (30.2 ps). The TDONN chip has successfully implemented four-class classification in different modalities (vision, audio, and touch) and achieve 85.7% accuracy on multimodal test sets. Our work opens up a new avenue for multimodal deep learning with integrated photonic processors, providing a potential solution for low-power AI large models using photonic technology. Most photonic processors can only handle a single data modality due to the lack of abundant parameter training in optical domain. Here, authors propose and demonstrate a trainable diffractive optical neural network chip based on on-chip diffractive optics with tunable elements to address these constraints.

In emerging artificial intelligence applications, massive matrix operations require high computing speed and energy efficiency. Optical computing can realize high-speed parallel information processing with ultra-low energy consumption on photonic integrated platforms or in free space, which can well meet these domain-specific demands. In this review, we firstly introduce the principles of photonic matrix computing implemented by three mainstream schemes, and then review the research progress of optical neural networks (ONNs) based on photonic matrix computing. In addition, we discuss the advantages of optical computing architectures over electronic processors as well as current challenges of optical computing and highlight some promising prospects for the future development.

Magnetic resonance (MR) angiography is one of the main diagnostic approaches for cardiac-cerebral vascular diseases. Nevertheless, the non-contrast-enhanced MR angiography suffers from its intrinsic problems derived from the blood flow-dependency, while the clinical Gd-chelating contrast agents are limited by their rapid vascular extravasation. Herein, we report a hypersensitive MR angiography strategy based on interlocking stratagem of zwitterionic Gd-chelate contrast agents (PAA-Gd). The longitudinal molar relaxivity of PAA-Gd was 4.6-times higher than that of individual Gd-chelates as well as appropriate blood half-life (73.8 min) and low immunogenicity, enabling sophisticated micro-vessels angiography with a resolution at the order of hundred micrometers. A series of animal models of cardiac-cerebrovascular diseases have been built for imaging studies on a 7.0 T MRI scanner, while the clinical translation potential of PAA-Gd has been evaluated on swine on a 3.0 T clinical MRI scanner. The current studies offer a promising strategy for precise diagnosis of vascular diseases. Current contrast-enhanced magnetic resonance angiography approaches are sub-optimal. Here the authors present a hypersensitive MR angiography strategy based on interlocking stratagem of zwitterionic Gd-chelate contrast agents (PAA-Gd), enabling sophisticated micro-vessel angiography of cardiac-cerebrovascular diseases with ultrahigh resolution.

In this paper, we prove that the isospectral flows associated with both the x-part and the n-part of the Lax pair of the semi-discrete lattice potential Korteweg–de Vries equation are symmetries of the equation. Furthermore, we show that these two hierarchies of symmetries are equivalent. Additionally, we construct the non-isospectral flows associated with the x-part of the Lax pair, which can be interpreted as the master symmetries of the semi-discrete lattice potential Korteweg–de Vries equation.

Central nervous system (CNS) diseases have been widely acknowledged as one of the major healthy concerns globally, which lead to serious impacts on human health. There will be about 135 million CNS diseases cases worldwide by mid-century, and CNS diseases will become the second leading cause of death after the cardiovascular disease by 2040. Most CNS diseases lack of effective diagnostic and therapeutic strategies with one of the reasons that the biological barrier extremely hampers the delivery of theranostic agents. In recent years, nanotechnology-based drug delivery is a quite promising way for CNS diseases due to excellent properties. Among them, cell membrane-based nanomaterials with natural bio-surface, high biocompatibility and biosafety, are of great significance in both the diagnosis and treatment of different CNS diseases. In this review, the state of art of the fabrication of cell membranes-based nanomaterials is introduced. The characteristics of different CNS diseases, and the application of cell membranes-based nanomaterials in the theranostics are summarized. In addition, the future prospects and limitations of cell membrane nanotechnology are anticipated. Through summarizing the state of art of the fabrication, giving examples of CNS diseases, and highlighting the applications in theranostics, the current review provides designing methods and ideas for subsequent cell membrane nanomaterials.

Explainable machine learning methods with a specific mathematical model provide insights into how the model works. We propose a new mode that contains a two-layer architecture for hyperspectral image (HSI) classification. In the front-end learning layer, superpixel segmentation and mathematical models are combined to achieve the band selection, which obtains the data re-expression in a lower dimension. The mathematical model uses the l2,1 norm and graph regularized term, which helps induce sparsity, improve robustness to outliers and noise, and enhance the explainability of the data re-expression. We employ the support vector machine or the K-nearest neighbor algorithms in the back-end layer to classify low-dimensional data. Finally, the two-layer mode classification method is applied to the three real HSI dataset classifications. Numerical results show that the overall classification accuracy of our method is improved.

Computing with light is widely recognized as a promising paradigm for overcoming the energy and latency limitations of electronic computing. However, the energy consumption and latency in current optical computing hardware predominantly arise in the electrical domain rather than the optical domain, primarily due to frequent signal conversions between optical (analog) and electrical (digital) formats. Furthermore, as the operating frequency of optical computing surpasses the GHz range, the synchronization of parallel electrical signals and the management of optical delays become increasingly critical. These challenges exacerbate energy consumption and latency, particularly in recurrent optical operations. To address these limitations, we propose a novel asynchronous computing paradigm for on-chip optical recurrent accelerators based on wavelength encoding, effectively mitigating synchronization challenges. By leveraging the intrinsic causality of wavelength relay, our approach eliminates the need for rigorous temporal alignment. To demonstrate the flexibility and efficacy of this asynchronous paradigm, we present two advanced recurrent models—an optical hidden Markov model and an optical recurrent neural network—monolithically integrated for the first time. These models incorporate hundreds of linear and nonlinear computing units densely packed into a compact footprint of just 10 mm 2 . Experimental evaluations on various benchmark tasks underscore the superior energy efficiency and low latency of the proposed asynchronous optical accelerators. This innovation enables the efficient processing of large-scale parallel signals and positions optical processors as a pivotal technology for applications such as autonomous driving and intelligent robotics.

State-of-the-art deep learning models can converse and interact with humans by understanding their emotions, but the exponential increase in model parameters has triggered an unprecedented demand for fast and low-power computing. Here, we propose a microcomb-enabled integrated optical neural network (MIONN) to perform the intelligent task of human emotion recognition at the speed of light and with low power consumption. Large-scale tensor data can be independently encoded in dozens of frequency channels generated by the on-chip microcomb and computed in parallel when flowing through the microring weight bank. To validate the proposed MIONN, we fabricated proof-of-concept chips and a prototype photonic-electronic artificial intelligence (AI) computing engine with a potential throughput up to 51.2 TOPS (tera-operations per second). We developed automatic feedback control procedures to ensure the stability and 8 bits weighting precision of the MIONN. The MIONN has successfully recognized six basic human emotions, and achieved 78.5 % accuracy on the blind test set. The proposed MIONN provides a high-speed and energy-efficient neuromorphic computing hardware for deep learning models with emotional interaction capabilities.

In recent years, shape memory alloys (SMAs) have been applied in the vibration control of engineering structures due to their special properties such as super elasticity and high damping, and the study of the performance of SMA wires has been relatively comprehensive, while research on the fatigue performance of SMA bars via cyclic tensile tests has been pretty rare, and low-cycle fatigue test has not been reported. However, the damage to building structures caused by earthquakes is of high-strain, low-cycle fatigue; therefore, in order for SMA bars to be used in seismic design, low-cycle fatigue tests were conducted on SMA bars with a diameter of 14 mm in this paper. Firstly, specimens were heat treated at a constant temperature of 350 °C for 30 min; other specimens were heat treated at a constant temperature of 400 °C for 15 min, while the rests were heat treated under a constant temperature of 400 °C for 30 min. Secondly, the energy dissipation capacity and residual strain of the SMA bar specimens were determined using the low-cycle fatigue test, in which the strain amplitudes were 2.5%, 3.5% and 3.75%. Additionally, the stress–strain relationship for SMA bars under cyclic loading was given. Finally, low-cycle fatigue properties of SMA bars were numerically simulated in the comparison analysis with the experimental results to verify their feasibility. Thus, it is proved that SMA bars can be recommend for seismic design building structures.