Quantization Levels Research Articles

Hardware assurance with silicon photonic physical unclonable functions

In the modern landscape of optical communication networks, ensuring robust security is increasingly critical, particularly for applications requiring seamless integration and minimal attack surfaces. Photonic Physical Unclonable Functions (PUFs) leverage the response from the photonic devices that are prone to inherent physical variations to generate unique and unpredictable signature identifiers which are then utilized by an authentication system for identification or encryption purposes. These photonic PUFs can be cohesively integrated into systems that use optical communication, whereas using electronic PUFs would introduce additional vulnerabilities due to the need for signal-domain conversions between optical and electronic signals. In this paper, we present the design, fabrication, and experimental evaluation of advanced silicon-photonic-based PUFs utilizing Contra-Directional Coupler (CDC) structures. These structures offer a complex design space and are intrinsically sensitive to fabrication-process variations, making them ideal for creating unique and secure responses. We introduce several innovative design enhancements, including randomized corrugation functions, perforated designs, and ring-assisted CDCs, to increase the complexity and unpredictability of the CDC response. Measurement results from the fabricated CDCs demonstrate their capability to achieve an average Hamming distance threshold of over 0.2, effectively distinguishing between legitimate devices and their copies. We rigorously tested these fabricated designs against three different machine-learning-based attack scenarios. The results showed a Hamming distance of over 0.4 with a standard deviation of less than 0.01 at a quantization level of three, using 10,000 samples of challenge-response pairs. These findings underscore the potential of silicon photonic PUFs in enhancing security for optical communication systems of different scales. The integration of such photonic PUFs offers robust and reliable security solutions for applications where traditional electronic methods introduce additional attack surfaces and fail to provide adequate protection.

Electrocardiogram Signal Compression Using Deep Convolutional Autoencoder with Constant Error and Flexible Compression Rate

ObjectivesElectrocardiogram (ECG) signals are beneficial for diagnosing cardiac diseases. The cardiac patients' life quality likely increases with continuous or long-period recording and monitoring of ECG signals, leading to better and early diagnosis of disease and heart attacks. However, continuous ECG recording necessitates high data rates and storage, which means high costs. Therefore, ECG compression is a handy concept that facilitates continuous monitoring of ECG signals. Deep neural networks open up new horizons for compression and also for ECG compression by providing high compression rates and quality. Although they bring constant compression ratios with better average quality, the compression quality of individual samples is not guaranteed, which may lead to misdiagnoses. This study aims to investigate the effect of compression quality on the diagnoses and to develop a deep neural network-based compression strategy that guarantees a quality-bound in return for varying compression ratios.Materials and methodsThe effect of the compression quality on the arrhythmia diagnoses is tested by comparing the performance of the deep learning-based ECG classifier on the original ECG recordings and the distorted recordings using a lossy compression algorithm with different compression error levels. Then, a compression error upper limit is calculated in terms of normalized percent root mean square difference (PRDN) error, which also coincides with the findings of the previous studies in the literature. Lastly, to enable deep learning in ECG compression, a single encoder-multi-decoder convolutional autoencoder architecture, and multiple quantization levels are proposed to guarantee a desired upper limit on the error rate.ResultsThe efficiency of the proposed method is demonstrated on a popular benchmark data set for ECG compression methods using a transfer learning approach. The PRDN error is fixed to various values, and the average compression rates are reported. An average of 13.019:1 compression is achieved for a 10% PRDN error rate, assessed as a fair quality threshold for reconstruction error. It has also been shown that the compression model has a runtime that can be run in real-time on wearable devices such as commercial smartwatches.ConclusionThis study proposes a deep learning-based ECG compression algorithm that guarantees a desired upper limit on the compression error. This model may facilitate an eHealth solution for continuous monitoring of ECG signals of individuals, especially cardiac patients.

DLAS: A Conceptual Model for Across-Stack Deep Learning Acceleration

Deep Neural Networks (DNNs) are very computationally demanding, which presents a significant barrier to their deployment, especially on resource-constrained devices. Significant work from both the machine learning and computing systems communities has attempted to accelerate DNNs. However, the number of techniques available and the required domain knowledge for their exploration continues to grow, making design space exploration (DSE) increasingly difficult. To unify the perspectives from these two communities, this paper introduces the Deep Learning Acceleration Stack (DLAS), a conceptual model for DNN deployment and acceleration. We adopt a six-layer representation that organizes and illustrates the key areas for DNN acceleration, from machine learning to software and computer architecture. We argue that the DLAS model balances simplicity and expressiveness, assisting practitioners from various domains in tackling co-design acceleration challenges. We demonstrate the interdependence of the DLAS layers, and thus the need for co-design, through an across-stack perturbation study, using a modified tensor compiler to generate experiments for combinations of a few parameters across the DLAS layers. Our perturbation study assesses the impact on inference time and accuracy when varying DLAS parameters across two datasets, seven popular DNN architectures, four compression techniques, three algorithmic primitives (with sparse and dense variants), untuned and auto-scheduled code generation, and four hardware platforms. The study observes significant changes in the relative performance of design choices with the introduction of new DLAS parameters (e.g., the fastest algorithmic primitive varies with the level of quantization). Given the strong evidence for the need for co-design, and the high costs of DSE, DLAS offers a valuable conceptual model for better exploring advanced co-designed accelerated deep learning solutions.

The digital pheromone: Building digital identity of smartphone users based on time-varying multivariates

The increasing reliance on smartphones necessitates secure, accurate user identification methods. This study introduces “digital pheromones”, a novel digital identity creation method based on smartphone user behaviors. Represented by a 56-bit binary number, this identity is derived from quantization and concatenation of 14 time-varying smartphone variables. Motivated by the need for enhanced security and privacy, the study employs statistical analysis to measure user behavior dispersion, showing that wider dispersion variables facilitate faster user uniqueness at lower quantization levels. The results demonstrate 100% uniqueness, ensuring no two users share the same 56-bit identity. This research offers a robust framework for future advancements in smartphone user authentication and identification systems.

Non-Destructive Monitoring of Peanut Leaf Area Index by Combing UAV Spectral and Textural Characteristics

The leaf area index (LAI) is a crucial metric for indicating crop development in the field, essential for both research and the practical implementation of precision agriculture. Unmanned aerial vehicles (UAVs) are widely used for monitoring crop growth due to their rapid, repetitive capture ability and cost-effectiveness. Therefore, we developed a non-destructive monitoring method for peanut LAI, combining UAV vegetation indices (VI) and texture features (TF). Field experiments were conducted to capture multispectral imagery of peanut crops. Based on these data, an optimal regression model was constructed to estimate LAI. The initial computation involves determining the potential spectral and textural characteristics. Subsequently, a comprehensive correlation study between these features and peanut LAI is conducted using Pearson’s product component correlation and recursive feature elimination. Six regression models, including univariate linear regression, support vector regression, ridge regression, decision tree regression, partial least squares regression, and random forest regression, are used to determine the optimal LAI estimation. The following results are observed: (1) Vegetation indices exhibit greater correlation with LAI than texture characteristics. (2) The choice of GLCM parameters for texture features impacts estimation accuracy. Generally, smaller moving window sizes and higher grayscale quantization levels yield more accurate peanut LAI estimations. (3) The SVR model using both VI and TF offers the utmost precision, significantly improving accuracy (R2 = 0.867, RMSE = 0.491). Combining VI and TF enhances LAI estimation by 0.055 (VI) and 0.541 (TF), reducing RMSE by 0.093 (VI) and 0.616 (TF). The findings highlight the significant improvement in peanut LAI estimation accuracy achieved by integrating spectral and textural characteristics with appropriate parameters. These insights offer valuable guidance for monitoring peanut growth.

Antiwindup Compensation for Unstable Rigid-Body Systems with Quantized and Saturated Inputs

It is well known that actuator saturation can cause destabilization and degradation in performance; similar problems are encountered when actuation is quantized. This study proposes the design of an antiwindup compensator for systems with actuators that are limited to a finite number of quantization levels. This combination of discrete-level actuation and saturation poses a unique antiwindup problem that has not yet been solved. To surmount this combined issue, an antiwindup compensator is proposed, which provides ultimate boundedness of the system state within a prescribed region and guarantees that the state does not stray outside a larger compact set. The use of shifted ramp functions enables a less conservative bound on the control-signal error, which yields significantly lower L2 gain bounds compared to a standard sector-bound antiwindup design approach. A numerical simulation example illustrates the effectiveness on a rigid-body system, which inspired this study.

Novel quantized iterative learning control based on spherical polar coordinates

AbstractThis study investigates the iterative learning control method for discrete‐time systems with data quantization, and it employs a quantizer based on spherical polar coordinates. The quantizer uses spherical polar coordinates to transform an unquantized signal into one with a predetermined quantization level. The quantization capability is dependent on the size of the support sphere, which is dynamically adjustable. In the scenarios of system output quantization, general quantization, and tracking error quantization, the learning control schemes are developed with consideration of the smallest sphere quantization features. The radii of the support spheres are designed independently based on the characteristics of the quantized signals (system output, control input, tracking error). The findings suggest that the system can achieve a bounded error convergence in the scenarios of system output and general quantization. In the context of quantization in tracking error, it is achievable to attain tracking performance with zero error. The example of a permanent magnet synchronous motor is provided to illustrate the findings of the study.

Effect of the One-to-Many Relationship between the Depth and Spectral Profile on Shallow Water Depth Inversion Based on Sentinel-2 Data

In shallow water, Sentinel-2 multispectral imagery has only four visible bands and limited quantization levels, which easily leads to the occurrence of the same spectral profile but different depth (SSPBDD) phenomenon, resulting in a one-to-many relationship between water depth and spectral profile. Investigating the impact of this relationship on water depth inversion models is the main objective of this paper. The Stumpf model and three machine learning models (Random Forest, Support Vector Machine, and Mixture Density Network) are employed, and the performance of these models is analysed based on the spatial distribution of the training dataset and the input information composition of these models. The results show that the root mean square errors (RMSEs) of the depth inversion of Random Forest and Support Vector Machine are significantly affected by the spatial distribution of the training dataset, while minimal effects are observed for the Stumpf model and the Mixture Density Network model. The SSPBDD phenomenon is widespread in Sentinel-2 images at all depths, particularly between 5 m and 15 m, with most of the depth maximum difference of the SSPBDD pixels ranging from 0 to 5 m. The SSPBDDs phenomenon can significantly reduce the inversion accuracy of any model. The number and the depth maximum difference of the SSPBDDs pixels are the main influencing factors. However, by increasing the visible spectral information and the spatial neighbourhood information in the input layer of machine learning models, the inversion accuracy and stability of the models can be improved to a certain extent. Among the models, the Mixture Density Network achieves the best inversion accuracy and stability.

Magnetism and electronic properties of ConMoP (n = 1 ~ 5) cluster: a DFT study.

Hydrogen has emerged as a promising clean energy carrier, underscoring the imperative need to comprehend its adsorption mechanisms. This study delves into the magnetic and electronic properties of Co-Mo-P clusters, aiming to unveil their catalytic potential in hydrogen production. Employing density functional theory (DFT), we optimized cluster configurations and scrutinized their magnetic behaviors. Our investigation unveiled 16 stable configurations of the ConMoP (n = 1 ~ 5) cluster, predominantly in steric forms. The magnetic attributes were primarily ascribed to the d orbitals of Co metal atoms, with Co3MoP exhibiting exceptional magnetic characteristics. Analysis of density of state diagrams revealed the prevalence of spin-up α-electrons in d orbitals, while spin-down β-electrons attenuated overall magnetic properties. Localized orbital (LOL) analysis highlighted stable covalent bonds within the clusters, affirming their catalytic potential. Orbital delocalization index (ODI) analysis revealed diverse spatial distribution ranges for orbitals across different configurations, suggesting a progressive attenuation of off-domain properties with increasing cluster size. Furthermore, infrared spectroscopy unveiled distinct vibrational peaks in various configurations, indicative of unique infrared activities. These findings contribute to a nuanced theoretical understanding of Co-Mo-P clusters and pave the path for future research aimed at augmenting their catalytic efficiency in hydrogen production. This study underscores the viability of Co-Mo-P clusters as alternatives to conventional Pt catalysts, offering insights into the design of novel materials for sustainable energy applications. Further research is warranted to explore the behavior of the Co-Mo-P system under diverse reaction conditions, fostering advancements in materials and energy science. In this study, we harnessed the ConMoP (n = 1 ~ 5) cluster as a simulation platform for probing the local structure of the material. Our aim was to scrutinize the magnetism, electronic characteristics influenced by the varying metal atoms within these clusters. A systematic exploration involved incrementing the number of metal atoms and expanding the cluster size to elucidate the corresponding property variations. Density functional theory (DFT) calculations were pivotal to our methodology, employing the B3LYP hybrid functional implemented in the Gaussian16 software package. The ConMoP (n = 1 ~ 5) cluster underwent optimization calculations and vibrational analysis at the def2-tzvp quantization level, yielding optimized configurations with diverse spin multiplet degrees. To comprehensively characterize and visually represent the stability, electronic features, and catalytic attributes of these configurations, we employed a suite of computational tools. Specifically, quantum chemistry software GaussView and wave function analysis software Multiwfn played integral roles. Through the integrated use of these computational tools, we acquired valuable insights into the magnetism, electronic characteristics of the ConMoP (n = 1 ~ 5) cluster, shedding light on their dependency on distinct metal atoms.

Evaluation of the effects of K-Means Clustered-Based Weight Quantization on a Keras Library Based Convolutional Neural Network for Hand Written Digit Image Recognition

A lot of Convolutional Neural Networks (CNNs) have been implemented using FPGAs for the past years. Subsequently, memory saving features were added to the CNN through weight quantization using K-means clustering. A future goal on an ASIC design, involving CNN and weight quantization working together in one chip, can give way to an automated procedure of memory-saving CNN design. In this paper an evaluation was done on the effect of quantizing the weights of a Keras library-based CNN using K means clustering. Various values of K in K-means clustering were tested to see its effects on the CNN accuracy performance. This paper presents first the design approach of a Keras library based Convolutional Neural Network (CNN) for hand-written digit images. It then presents a hardware model design of K-Means clustering algorithm using VHDL. The performance of CNN for image recognition was then tested for various levels of weight quantization using K-means clustering algorithm. Simulation results showed a compression of weights as high as 60% resulted to less than 1% reduction in CNN’s accuracy. The findings in this paper will serve as guide in determining the relevant values of K i.e. the compression ratio, for future ASIC design on this topic.

MODELING OF ANALOG-TO-DIGITAL SIGNAL CONVERTERS FOR SENSOR MICROSYSTEMS IN THE MICROWIND SOFTWARE

For the results of simulating the operation of analog-digital converters, we chose such a converter as the serial approximation ADC (SAR ADC). A sequential approximation ADC is a type of analog-to-digital converter that converts a continuous analog waveform to a discrete digital representation by binary search through all possible quantization levels before finally converging on a digital output for each conversion. There are three important blocks in the SAR ADC architecture: the sample and hold circuit (Sample and Hold, S/H), the comparator, and the SAR block. The topology and principle of operation of which was modeled in the Micro Wind software environment.

Review of neuromorphic computing based on NAND flash memory.

The proliferation of data has facilitated global accessibility, which demands escalating amounts of power for data storage and processing purposes. In recent years, there has been a rise in research in the field of neuromorphic electronics, which draws inspiration from biological neurons and synapses. These electronics possess the ability to perform in-memory computing, which helps alleviate the limitations imposed by the 'von Neumann bottleneck' that exists between the memory and processor in the traditional von Neumann architecture. By leveraging their multi-bit non-volatility, characteristics that mimic biology, and Kirchhoff's law, neuromorphic electronics offer a promising solution to reduce the power consumption in processing vector-matrix multiplication tasks. Among all the existing nonvolatile memory technologies, NAND flash memory is one of the most competitive integrated solutions for the storage of large volumes of data. This work provides a comprehensive overview of the recent developments in neuromorphic computing based on NAND flash memory. Neuromorphic architectures using NAND flash memory for off-chip learning are presented with various quantization levels of input and weight. Next, neuromorphic architectures for on-chip learning are presented using standard backpropagation and feedback alignment algorithms. The array architecture, operation scheme, and electrical characteristics of NAND flash memory are discussed with a focus on the use of NAND flash memory in various neural network structures. Furthermore, the discrepancy of array architecture between on-chip learning and off-chip learning is addressed. This review article provides a foundation for understanding the neuromorphic computing based on the NAND flash memory and methods to utilize it based on application requirements.

Non-Serial Quantization-Aware Deep Optics for Snapshot Hyperspectral Imaging.

Deep optics has been endeavoring to capture hyperspectral images of dynamic scenes, where the optical encoder plays an essential role in deciding the imaging performance. Our key insight is that the optical encoder of a deep optics system is expected to keep fabrication-friendliness and decoder-friendliness, to be faithfully realized in the implementation phase and fully interacted with the decoder in the design phase, respectively. In this paper, we propose the non-serial quantization-aware deep optics (NSQDO), which consists of the fabrication-friendly quantization-aware model (QAM) and the decoder-friendly non-serial manner (NSM). The QAM integrates the quantization process into the optimization and adaptively adjusts the physical height of each quantization level, reducing the deviation of the physical encoder from the numerical simulation through the awareness of and adaptation to the quantization operation of the DOE physical structure. The NSM bridges the encoder and the decoder with full interaction through bidirectional hint connections and flexibilize the connections with a gating mechanism, boosting the power of joint optimization in deep optics. The proposed NSQDO improves the fabrication-friendliness and decoder-friendliness of the encoder and develops the deep optics framework to be more practical and powerful. Extensive synthetic simulation and real hardware experiments demonstrate the superior performance of the proposed method.

Quantization of the Energy Spectrum of Holes in the Adiabatic Potential of the Electron

The spectra of the interband absorption of microscopic CdS crystals, 15 to 30 Å in size, grown in a transparent insulating matrix, are analyzed. A structural feature caused by quantization of the energy spectrum of the hole in the adiabatic potential of its Coulomb interaction with the electron has been detected near the transitions to the lower level of the size quantization of the electron.

Quantized iterative learning control for impulsive differential inclusion systems with data dropouts

This paper studies the quantized iterative learning control with encoding–decoding mechanism of a class of impulsive differential inclusion systems with random data dropouts. First, the set-valued mappings in the differential inclusion systems are transformed into single-valued mappings by using the Steiner-type selector. Then, a learning algorithm based on the intermittent update principle is designed to address the data asynchronism problem caused by two-sided data dropouts. If the data are successfully transmitted at the actuator and measurement sides, then the control input is effectively updated. Furthermore, a suitable scaling sequence is introduced to ensure the system output to achieve zero-error tracking performance for a desired trajectory. An upper bound of the quantization level is determined such that the quantization error is always bounded. The results show that the quantization method reduces the burden of network communication at the cost of increasing the amount of computation, and the learning algorithm does not require the data dropouts to satisfy a certain probability distribution. Finally, the effectiveness of the learning algorithm is verified by numerical simulations of the switched reluctance motor system.

Asymmetric particle-antiparticle Dirac equation: second quantization

We build the fully relativistic quantum field theory related to the asymmetric Dirac fields first presented in a prequel to this work. These fields are solutions of the asymmetric Dirac equation, a Lorentz covariant Dirac-like equation whose positive and ‘negative’ frequency plane wave solutions’ dispersion relations are no longer degenerate. At the second quantization level, we show that this implies that particles and antiparticles sharing the same wave number have different energies and momenta. In spite of that, we prove that by properly fixing the values of the relativistic invariants that define the asymmetric Dirac free field Lagrangian density, we can build a consistent, fully relativistic, and renormalizable quantum electrodynamics (QED) that is empirically equivalent to the standard QED. We discuss the reasons and implications of this non-trivial equivalence, exploring qualitatively other scenarios in which the asymmetric Dirac fields may lead to beyond the standard model predictions. We give a complete account of how the asymmetric Dirac fields and the corresponding annihilation and creation operators transform under improper Lorentz transformations (parity and time reversal operations) and under the charge conjugation operation. We also prove that the present theory respects the CPT theorem.

Attention Round for post-training quantization

Quantization methods for convolutional neural network models can be broadly categorized into post-training quantization (PTQ) and quantization aware training (QAT). While PTQ offers the advantage of requiring only a small portion of the data for quantization, the resulting quantized model may not be as effective as QAT. To address this limitation, this paper proposes a novel quantization function named Attention Round. Unlike traditional quantization function that map 32 bit floating-point value w to nearby quantization levels, Attention Round allows w to be mapped to all possible quantization levels in the entire quantization space, expanding the quantization optimization space. The possibilities of mapping w to different quantization levels are inversely correlated with the distance between w and the quantization levels, regulated by a Gaussian decay function. Furthermore, to tackle the challenge of mixed precision quantization, this paper introduces a lossy coding length measure to assign quantization precision to different layers of the model, eliminating the need for solving a combinatorial optimization problem. Experimental evaluations on various models demonstrate the effectiveness of the proposed method. Notably, for ResNet18 and MobileNetV2, the PTQ approach achieves comparable quantization performance to QAT while utilizing only 1024 training data and 10 min for the quantization process.

Connectivity comparison of uniform quantization-based graph transformation and its application in spectrum sensing

In the uniform quantization-based graph transformation of random sequences, the connectivity of the graph is an important feature and can be considered as a basis for designing graph domain processing algorithms. However, the current research on graph transformation focuses mostly on analysing the majorization orders of different input distributions to achieve a comparison of graph connectivity. Generally, calculating majorization order between two candidate probability density functions requires two inversions and one integration for each probability density function, and obtaining an analytical solution is extremely difficult. Even if an analytical solution can be obtained, considering the impact of function transformations such as normalization, analysing the reservation of majorization order is challenging. However, the dispersive order calculation only involves one inversion and differentiation of the distribution functions, and it implies majorization under certain conditions. Consequently, the objectives of this study were to investigate the dispersive order and its order reservation between two random sequences with different probability distributions and to compare the differences in connectivity between the graphs generated under the same conditions, such as the number of uniform quantization levels and sample size. In addition, the properties of complete graphs generated from random sequences were evaluated, including the influence of the probability vector of vertices on the graphs, the sample size and number of vertices on the probability of generating complete graphs. As an application, a graph-based spectrum sensing algorithm based on the block-maxima of the power spectrum was formulated. Extensive simulations showed that the proposed algorithm outperforms the existing graph-based sensing algorithms both in detection performance and the computational complexity.

Distributed constrained optimization with periodic dynamic quantization

This paper uses the mirror descent algorithm with periodic dynamic quantization to solve constrained distributed optimization problems with limited communication channels. Due to the imperfect network environment, obtaining accurate information is impractical, and thus a communication scheme under quantization needs to be considered. A periodic dynamic quantizer with finite quantization levels is proposed in this paper to achieve exact optimization. Moreover, a time-varying control parameter in the mirror descent algorithm is designed to control the quantization error. After a comprehensive analysis, the proposed algorithm can obtain an optimal value, and the optimal convergence rate is O(1/T0.25).

Possibility of Exciton Bose-Einstein Condensation in CdSe Nanoplatelets.

The quasi-two-dimensional exciton subsystem in CdSe nanoplatelets is considered. It is theoretically shown that Bose-Einstein condensation (BEC) of excitons is possible at a nonzero temperature in the approximation of an ideal Bose gas and in the presence of an "energy gap" between the ground and the first excited states of the two-dimensional exciton center of inertia of the translational motion. The condensation temperature (Tc) increases with the width of the "gap" between the ground and the first excited levels of size quantization. It is shown that when the screening effect of free electrons and holes on bound excitons is considered, the BEC temperature of the exciton subsystem increases as compared to the case where this effect is absent. The energy spectrum of the exciton condensate in a CdSe nanoplate is calculated within the framework of the weakly nonideal Bose gas approximation, considering the specifics of two-dimensional Born scattering.