Activation Redistribution Based Hybrid Asymmetric Quantization Method of Neural Networks

The demand for adopting neural networks in resource-constrained embedded devices is continuously increasing.Quantization is one of the most promising solutions to reduce computational cost and memory storage on embedded devices.In order to reduce the complexity and overhead of deploying neural networks on Integeronly hardware,most current quantization methods use a symmetric quantization mapping strategy to quantize a floating-point neural network into an integer network.However,although symmetric quantization has the advantage of easier implementation,it is sub-optimal for cases where the range could be skewed and not symmetric.This often comes at the cost of lower accuracy.This paper proposed an activation redistribution-based hybrid asymmetric quantizationmethod for neural networks.The proposedmethod takes data distribution into consideration and can resolve the contradiction between the quantization accuracy and the ease of implementation,balance the trade-off between clipping range and quantization resolution,and thus improve the accuracy of the quantized neural network.The experimental results indicate that the accuracy of the proposed method is 2.02%and 5.52%higher than the traditional symmetric quantization method for classification and detection tasks,respectively.The proposed method paves the way for computationally intensive neural network models to be deployed on devices with limited computing resources.Codes will be available on https://github.com/ycjcy/Hybrid-Asymmetric-Quantization.

Reinforcement Learning Based Quantization Strategy Optimal Assignment Algorithm for Mixed Precision

The quantization algorithm compresses the original network by reducing the numerical bit width of the model,which improves the computation speed. Because different layers have different redundancy and sensitivity to databit width. Reducing the data bit width will result in a loss of accuracy. Therefore, it is difficult to determinethe optimal bit width for different parts of the network with guaranteed accuracy. Mixed precision quantizationcan effectively reduce the amount of computation while keeping the model accuracy basically unchanged. In thispaper, a hardware-aware mixed precision quantization strategy optimal assignment algorithm adapted to low bitwidth is proposed, and reinforcement learning is used to automatically predict the mixed precision that meets theconstraints of hardware resources. In the state-space design, the standard deviation of weights is used to measurethe distribution difference of data, the execution speed feedback of simulated neural network accelerator inferenceis used as the environment to limit the action space of the agent, and the accuracy of the quantization model afterretraining is used as the reward function to guide the agent to carry out deep reinforcement learning training. Theexperimental results show that the proposed method obtains a suitable model layer-by-layer quantization strategyunder the condition that the computational resources are satisfied, and themodel accuracy is effectively improved.The proposed method has strong intelligence and certain universality and has strong application potential in thefield of mixed precision quantization and embedded neural network model deployment.

In situ calibrated angle between the quantization axis and the propagating direction of the light field for trapping neutral atoms

The recently developed magic-intensity trapping technique of neutral atoms efficiently mitigates the detrimental effect of light shifts on atomic qubits and substantially enhances the coherence time. This technique relies on applying a bias magnetic field precisely parallel to the wave vector of a circularly polarized trapping laser field. However, due to the presence of the vector light shift experienced by the trapped atoms, it is challenging to precisely define a parallel magnetic field, especially at a low bias magnetic field strength, for the magic-intensity trapping of85Rb qubits. In this work, we present a method to calibrate the angle between the bias magnetic field and the trapping laser field with the compensating magnetic fields in the other two directions orthogonal to the bias magnetic field direction. Experimentally, with a constantdepth trap and a fixed bias magnetic field, we measure the respective resonant frequencies of the atomic qubits in a linearly polarized trap and a circularly polarized one via the conventional microwave Rabi spectra with different compensating magnetic fields and obtain the corresponding total magnetic fields via the respective resonant frequencies using the Breit–Rabi formula. With known total magnetic fields, the angle is a function of the other two compensating magnetic fields.Finally, the projection value of the angle on either of the directions orthogonal to the bias magnetic field direction can be reduced to 0(4)° by applying specific compensating magnetic fields. The measurement error is mainly attributed to the fluctuation of atomic temperature. Moreover, it also demonstrates that, even for a small angle, the effect is strong enough to cause large decoherence of Rabi oscillation in a magic-intensity trap. Although the compensation method demonstrated here is explored for the magic-intensity trapping technique, it can be applied to a variety of similar precision measurements with trapped neutral atoms.

A Novel Quantization and Model Compression Approach for Hardware Accelerators in Edge Computing

Massive computational complexity and memory requirement of artificial intelligence models impede their deploy-ability on edge computing devices of the Internet of Things(IoT).While Power-of-Two(PoT)quantization is pro-posed to improve the efficiency for edge inference of Deep Neural Networks(DNNs),existing PoT schemes require a huge amount of bit-wise manipulation and have large memory overhead,and their efficiency is bounded by the bottleneck of computation latency and memory footprint.To tackle this challenge,we present an efficient inference approach on the basis of PoT quantization and model compression.An integer-only scalar PoT quantization(IOS-PoT)is designed jointly with a distribution loss regularizer,wherein the regularizer minimizes quantization errors and training disturbances.Additionally,two-stage model compression is developed to effectively reduce memory requirement,and alleviate bandwidth usage in communications of networked heterogenous learning systems.The product look-up table(P-LUT)inference scheme is leveraged to replace bit-shifting with only indexing and addition operations for achieving low-latency computation and implementing efficient edge accelerators.Finally,comprehensive experiments on Residual Networks(ResNets)and efficient architectures with Canadian Institute for Advanced Research(CIFAR),ImageNet,and Real-world Affective Faces Database(RAF-DB)datasets,indicate that our approach achieves 2×∼10×improvement in the reduction of both weight size and computation cost in comparison to state-of-the-art methods.A P-LUT accelerator prototype is implemented on the Xilinx KV260 Field Programmable Gate Array(FPGA)platform for accelerating convolution operations,with performance results showing that P-LUT reduces memory footprint by 1.45×,achieves more than 3×power efficiency and 2×resource efficiency,compared to the conventional bit-shifting scheme.

Image Steganography by Pixel-Value Differencing Using General Quantization Ranges

A new steganographic method by pixel-value differencing(PVD)using general quantization ranges of pixel pairs’difference values is proposed.The objective of this method is to provide a data embedding technique with a range table with range widths not limited to powers of 2,extending PVD-based methods to enhance their flexibility and data-embedding rates without changing their capabilities to resist security attacks.Specifically,the conventional PVD technique partitions a grayscale image into 1×2 non-overlapping blocks.The entire range[0,255]of all possible absolute values of the pixel pairs’grayscale differences in the blocks is divided into multiple quantization ranges.The width of each quantization range is a power of two to facilitate the direct embedding of the bit information with high embedding rates.Without using power-of-two range widths,the embedding rates can drop using conventional embedding techniques.In contrast,the proposed method uses general quantization range widths,and a multiple-based number conversion mechanism is employed skillfully to implement the use of nonpower-of-two range widths,with each pixel pair being employed to embed a digit in the multiple-based number.All the message bits are converted into a big multiple-based number whose digits can be embedded into the pixel pairs with a higher embedding rate.Good experimental results showed the feasibility of the proposed method and its resistance to security attacks.In addition,implementation examples are provided,where the proposed method adopts non-power-of-two range widths and employsmultiple-based number conversion to expand the data-hiding and steganalysis-resisting capabilities of other PVD methods.

Bunch-length measurement at a bunch-by-bunch rate based on time–frequency-domain joint analysis techniques and its application

This paper presents a new technique for measuring the bunch length of a high-energy electron beam at a bunch-by-bunch rate in storage rings.This technique uses the time–frequency-domain joint analysis of the bunch signal to obtain bunch-by-bunch and turn-by-turn longitudinal parameters,such as bunch length and synchronous phase.The bunch signal is obtained using a button electrode with a bandwidth of several gigahertz.The data acquisition device was a high-speed digital oscilloscope with a sampling rate of more than 10 GS/s,and the single-shot sampling data buffer covered thousands of turns.The bunch-length and synchronous phase information were extracted via offline calculations using Python scripts.The calibration coefficient of the system was determined using a commercial streak camera.Moreover,this technique was tested on two different storage rings and successfully captured various longitudinal transient processes during the harmonic cavity debugging process at the Shanghai Synchrotron Radiation Facility(SSRF),and longitudinal instabilities were observed during the single-bunch accumulation process at Hefei Light Source(HLS).For Gaussian-distribution bunches,the uncertainty of the bunch phase obtained using this technique was better than 0.2 ps,and the bunch-length uncertainty was better than 1 ps.The dynamic range exceeded 10 ms.This technology is a powerful and versatile beam diagnostic tool that can be conveniently deployed in high-energy electron storage rings.

Learning Vector Quantization-Based Fuzzy Rules Oversampling Method

Imbalanced datasets are common in practical applications,and oversampling methods using fuzzy rules have been shown to enhance the classification performance of imbalanced data by taking into account the relationship between data attributes.However,the creation of fuzzy rules typically depends on expert knowledge,which may not fully leverage the label information in training data and may be subjective.To address this issue,a novel fuzzy rule oversampling approach is developed based on the learning vector quantization(LVQ)algorithm.In this method,the label information of the training data is utilized to determine the antecedent part of If-Then fuzzy rules by dynamically dividing attribute intervals using LVQ.Subsequently,fuzzy rules are generated and adjusted to calculate rule weights.The number of new samples to be synthesized for each rule is then computed,and samples from the minority class are synthesized based on the newly generated fuzzy rules.This results in the establishment of a fuzzy rule oversampling method based on LVQ.To evaluate the effectiveness of this method,comparative experiments are conducted on 12 publicly available imbalance datasets with five other sampling techniques in combination with the support function machine.The experimental results demonstrate that the proposed method can significantly enhance the classification algorithm across seven performance indicators,including a boost of 2.15%to 12.34%in Accuracy,6.11%to 27.06%in G-mean,and 4.69%to 18.78%in AUC.These show that the proposed method is capable of more efficiently improving the classification performance of imbalanced data.

A New Device for Gas-Liquid Flow Measurements Relying on Forced Annular Flow

A new measurement device,consisting of swirling blades and capsule-shaped throttling elements,is proposed in this study to eliminate typical measurement errors caused by complex flow patterns in gas-liquid flow.The swirling blades are used to transform the complex flow pattern into a forced annular flow.Drawing on the research of existing blockage flow meters and also exploiting the single-phase flow measurement theory,a formula is introduced to measure the phase-separated flow of gas and liquid.The formula requires the pressure ratio,Lockhart-Martinelli number(L-M number),and the gas phase Froude number.The unknown parameters appearing in the formula are fitted through numerical simulation using computational fluid dynamics(CFD),which involves a comprehensive analysis of the flow field inside the device from multiple perspectives,and takes into account the influence of pressure fluctuations.Finally,the measurement model is validated through an experimental error analysis.The results demonstrate that the measurement error can be maintained within±8%for various flow patterns,including stratified flow,bubble flow,and wave flow.

Enhancing Energy Efficiency with a Dynamic Trust Measurement Scheme in Power Distribution Network

The application of Intelligent Internet of Things(IIoT)in constructing distribution station areas strongly supports platform transformation,upgrade,and intelligent integration.The sensing layer of IIoT comprises the edge convergence layer and the end sensing layer,with the former using intelligent fusion terminals for real-time data collection and processing.However,the influx of multiple low-voltage in the smart grid raises higher demands for the performance,energy efficiency,and response speed of the substation fusion terminals.Simultaneously,it brings significant security risks to the entire distribution substation,posing a major challenge to the smart grid.In response to these challenges,a proposed dynamic and energy-efficient trust measurement scheme for smart grids aims to address these issues.The scheme begins by establishing a hierarchical trust measurement model,elucidating the trust relationships among smart IoT terminals.It then incorporates multidimensional measurement factors,encompassing static environmental factors,dynamic behaviors,and energy states.This comprehensive approach reduces the impact of subjective factors on trust measurements.Additionally,the scheme incorporates a detection process designed for identifying malicious low-voltage end sensing units,ensuring the prompt identification and elimination of any malicious terminals.This,in turn,enhances the security and reliability of the smart grid environment.The effectiveness of the proposed scheme in pinpointing malicious nodes has been demonstrated through simulation experiments.Notably,the scheme outperforms established trust metric models in terms of energy efficiency,showcasing its significant contribution to the field.

Measuring small longitudinal phase shifts via weak measurement amplification

Weak measurement amplification,which is considered as a very promising scheme in precision measurement,has been applied to various small physical quantities estimations.Since many physical quantities can be converted into phase signals,it is interesting and important to consider measuring small longitudinal phase shifts by using weak measurement.Here,we propose and experimentally demonstrate a novel weak measurement amplification-based small longitudinal phase estimation,which is suitable for polarization interferometry.We realize one order of magnitude amplification measurement of a small phase signal directly introduced by a liquid crystal variable retarder and show that it is robust to the imperfection of interference.Besides,we analyze the effect of magnification error which is never considered in the previous works,and find the constraint on the magnification.Our results may find important applications in high-precision measurements,e.g.,gravitational wave detection.

Improved spatio-temporal alignment measurement method for hull deformation

In this paper,an improved spatio-temporal alignment measurement method is presented to address the inertial matching measurement of hull deformation under the coexistence of time delay and large misalignment angle.Large misalignment angle and time delay often occur simultaneously and bring great challenges to the accurate measurement of hull deformation in space and time.The proposed method utilizes coarse alignment with large misalignment angle and time delay estimation of inertial measurement unit modeling to establish a brand-new spatiotemporal aligned hull deformation measurement model.In addition,two-step loop control is designed to ensure the accurate description of dynamic deformation angle and static deformation angle by the time-space alignment method of hull deformation.The experiments illustrate that the proposed method can effectively measure the hull deformation angle when time delay and large misalignment angle coexist.

Design and application of thickness measurement calibration system based on laser displacement sensor

This study aims to improve the accuracy and safety of steel plate thickness calibration.A differential noncontact thickness measurement calibration system based on laser displacement sensors was designed to address the problems of low precision of traditional contact thickness gauges and radiation risks of radiation-based thickness gauges.First,the measurement method and measurement structure of the thickness calibration system were introduced.Then,the hardware circuit of the thickness system was established based on the STM32 core chip.Finally,the system software was designed to implement system control to filter algorithms and human-computer interaction.Experiments have proven the excellent performance of the differential noncontact thickness measurement calibration system based on laser displacement sensors,which not only considerably improves measurement accuracy but also effectively reduces safety risks during the measurement process.The system offers guiding significance and application value in the field of steel plate production and processing.

A seismic elastic moduli module for the measurements of low-frequency wave dispersion and attenuation of fluid-saturated rocks under different pressures

Knowledge about the seismic elastic modulus dispersion,and associated attenuation,in fluid-saturated rocks is essential for better interpretation of seismic observations taken as part of hydrocarbon identification and time-lapse seismic surveillance of both conventional and unconventional reservoir and overburden performances.A Seismic Elastic Moduli Module has been developed,based on the forced-oscillations method,to experimentally investigate the frequency dependence of Young's modulus and Poisson's ratio,as well as the inferred attenuation,of cylindrical samples under different confining pressure conditions.Calibration with three standard samples showed that the measured elastic moduli were consistent with the published data,indicating that the new apparatus can operate reliably over a wide frequency range of f∈[1-2000,10^(6)]Hz.The Young's modulus and Poisson's ratio of the shale and the tight sandstone samples were measured under axial stress oscillations to assess the frequency-and pressure-dependent effects.Under dry condition,both samples appear to be nearly frequency independent,with weak pressure dependence for the shale and significant pressure dependence for the sandstone.In particular,it was found that the tight sandstone with complex pore microstructure exhibited apparent dispersion and attenuation under brine or glycerin saturation conditions,the levels of which were strongly influenced by the increased effective pressure.In addition,the measured Young's moduli results were compared with the theoretical predictions from a scaled poroelastic model with a reasonably good agreement,revealing that the combined fluid flow mechanisms at both mesoscopic and microscopic scales possibly responsible for the measured dispersion.

A fast forward computational method for nuclear measurement using volumetric detection constraints

Owing to the complex lithology of unconventional reservoirs,field interpreters usually need to provide a basis for interpretation using logging simulation models.Among the various detection tools that use nuclear sources,the detector response can reflect various types of information of the medium.The Monte Carlo method is one of the primary methods used to obtain nuclear detection responses in complex environments.However,this requires a computational process with extensive random sampling,consumes considerable resources,and does not provide real-time response results.Therefore,a novel fast forward computational method(FFCM)for nuclear measurement that uses volumetric detection constraints to rapidly calculate the detector response in various complex environments is proposed.First,the data library required for the FFCM is built by collecting the detection volume,detector counts,and flux sensitivity functions through a Monte Carlo simulation.Then,based on perturbation theory and the Rytov approximation,a model for the detector response is derived using the flux sensitivity function method and a one-group diffusion model.The environmental perturbation is constrained to optimize the model according to the tool structure and the impact of the formation and borehole within the effective detection volume.Finally,the method is applied to a neutron porosity tool for verification.In various complex simulation environments,the maximum relative error between the calculated porosity results of Monte Carlo and FFCM was 6.80%,with a rootmean-square error of 0.62 p.u.In field well applications,the formation porosity model obtained using FFCM was in good agreement with the model obtained by interpreters,which demonstrates the validity and accuracy of the proposed method.

Fast compressed sensing spectral measurement with adaptive gradient multiscale resolution

We propose a fast,adaptive multiscale resolution spectral measurement method based on compressed sensing.The method can apply variable measurement resolution over the entire spectral range to reduce the measurement time by over 75%compared to a global high-resolution measurement.Mimicking the characteristics of the human retina system,the resolution distribution follows the principle of gradually decreasing.The system allows the spectral peaks of interest to be captured dynamically or to be specified a priori by a user.The system was tested by measuring single and dual spectral peaks,and the results of spectral peaks are consistent with those of global high-resolution measurements.

Finesse measurement for high-power optical enhancement cavity

Finesse is a critical parameter for describing the characteristics of an optical enhancement cavity(OEC). This paper first presents a review of finesse measurement techniques, including a comparative analysis of the advantages, disadvantages, and potential limitations of several main methods from both theoretical and practical perspectives. A variant of the existing method called the free spectral range(FSR) modulation method is proposed and compared with three other finesse measurement methods, i.e., the fast-switching cavity ring-down(CRD) method, the rapidly swept-frequency(SF) CRD method, and the ringing effect method. A high-power OEC platform with a high finesse of approximately 16000 is built and measured with the four methods. The performance of these methods is compared, and the results show that the FSR modulation method and the fast-switching CRD method are more suitable and accurate than the other two methods for high-finesse OEC measurements. The CRD method and the ringing effect method can be implemented in open loop using simple equipment and are easy to perform. Additionally, recommendations for selecting finesse measurement methods under different conditions are proposed, which benefit the development of OEC and its applications.

Initial measurement of ultrafast charge exchange recombination spectroscopy on the EAST tokamak

Ultrafast charge exchange recombination spectroscopy(UF-CXRS)has been developed on the EAST tokamak(Yingying Li et al 2019 Fusion Eng.Des.146522)to measure fast evolutions of ion temperature and toroidal velocity.Here,we report the preliminary diagnostic measurements after relative sensitivity calibration.The measurement results show a much higher temporal resolution compared with conventional CXRS,benefiting from the usage of a prismcoupled,high-dispersion volume-phase holographic transmission grating and a high quantum efficiency,high-gain detector array.Utilizing the UF-CXRS diagnostic,the fast evolutions of the ion temperature and rotation velocity during a set of high-frequency small-amplitude edgelocalized modes(ELMs)are obtained on the EAST tokamak,which are then compared with the case of large-amplitude ELMs.

Micron-sized fiber diamond probe for quantum precision measurement of microwave magnetic field

We present a quantitative measurement of the horizontal component of the microwave magnetic field of a coplanar waveguide using a quantum diamond probe in fiber format.The measurement results are compared in detail with simulation,showing a good consistence.Further simulation shows fiber diamond probe brings negligible disturbance to the field under measurement compared to bulk diamond.This method will find important applications ranging from electromagnetic compatibility test and failure analysis of high frequency and high complexity integrated circuits.

Statistical Channel Modeling for Indoor VLC Communications Based on Channel Measurements

Visible light communication(VLC)has attracted much attention in the research of sixthgeneration(6G)systems.Furthermore,channel modeling is the foundation for designing efficient and robust VLC systems.In this paper,we present extensive VLC channel measurement campaigns in indoor environments,i.e.,an office and a corridor.Based on the measured data,the large-scale fading characteristics and multipath-related characteristics,including omnidirectional optical path loss(OPL),K-factor,power angular spectrum(PAS),angle spread(AS),and clustering characteristics,are analyzed and modeled through a statistical method.Based on the extracted statistics of the above-mentioned channel characteristics,we propose a statistical spatial channel model(SSCM)capable of modeling multipath in the spatial domain.Furthermore,the simulated statistics of the proposed model are compared with the measured statistics.For instance,in the office,the simulated path loss exponent(PLE)and the measured PLE are 1.96and 1.97,respectively.And,the simulated medians of AS and measured medians of AS are 25.94°and 24.84°,respectively.Generally,the fact that the simulated results fit well with measured results has demonstrated the accuracy of our SSCM.

Non-Kramers doublet ground state in a quaternary cubic compound PrRu_(2)In_(2)Zn_(18) investigated by ultrasonic measurements

We performed ultrasonic measurements on a quaternary cubic compound PrRu_(2)In_(2)Zn_(18) to explore the ground state properties derived from non-Kramers Γ_(3) doublet of Pr^(3+).PrRu_(2)In_(2)Zn_(18) is a quaternary derivative of the ternary compound PrRu_(2)Zn_(20) that exhibits a structural phase transition at T_S=138 K.In PrRu_(2)In_(2)Zn_(18),the Zn atoms at the 16c site in PrRu_(2)Zn_(20) are selectively replaced by In atoms.A monotonic increase was observed in the temperature dependence of elastic constants C_L=(C_(11)+2C_(12)+4C_(44))/3 and C_(T)=(C_(11)-C_(12)+C_(44))/3 in the temperature range around T_(S) to which an elastic softening was observed in(C_(11)-C_(12))/2 for PrRu_(2)Zn_(20).The disappearance of the softening indicates that the structural transition in PrRu_(2)Zn_(20) is suppressed by the substitution of Zn ions by In ones with a larger ionic radius.Alternatively,the C_(T) of PrRu_(2)In_(2)Zn_(18) exhibits a precursor Curie-type elastic softening toward low temperatures being responsible for the non-Kramers Γ_(3) ground state.We discuss the ground state and the evolution of the elastic properties of the different single-crystal samples of PrRu_(2)In_(2)Zn_(18) grown under different conditions.