On the basis of realization of beach information and its differentiating of high-resolution remote sensing image on coastal zone, extracting objects are carried through RS multi-scale diagnostic analysis, and fast inf...On the basis of realization of beach information and its differentiating of high-resolution remote sensing image on coastal zone, extracting objects are carried through RS multi-scale diagnostic analysis, and fast information extraction methods and key technologies are put forward. Meanwhile image segmentation methods are set forth for objects of coastal zone. And through the application of Otsu2D to the segmentation of water area and dock and the applying of Gabor filter to the separation and extraction of construction, some typical applications of high-resolution RS image are presented in the field of coastal zone surface objects＇ recognition. Quantizing high-resolution RS information on the coastal zone proved to be of great scientific and practical significance for coastal development and management.展开更多
A rule of how the ignition probability or invalidation probability of infantry grenade varies with the projectile quantity is presented based on the statistical analysis. The statistical induction, quantizing analysis...A rule of how the ignition probability or invalidation probability of infantry grenade varies with the projectile quantity is presented based on the statistical analysis. The statistical induction, quantizing analysis and data processing in the infantry grenade type approval test are completed, and we obtain how various test factors in the research item affect the test results(invalidation probability), i.e., quantized data. The acquisition of these quantized data provides theory basis and data auspice for further reasonable filtration of all test factors and the optimization of test scheme.展开更多
Exploring the expected quantizing scheme with suitable mixed-precision policy is the key to compress deep neural networks(DNNs)in high efficiency and accuracy.This exploration implies heavy workloads for domain expert...Exploring the expected quantizing scheme with suitable mixed-precision policy is the key to compress deep neural networks(DNNs)in high efficiency and accuracy.This exploration implies heavy workloads for domain experts,and an automatic compression method is needed.However,the huge search space of the automatic method introduces plenty of computing budgets that make the automatic process challenging to be applied in real scenarios.In this paper,we propose an end-to-end framework named AutoQNN,for automatically quantizing different layers utilizing different schemes and bitwidths without any human labor.AutoQNN can seek desirable quantizing schemes and mixed-precision policies for mainstream DNN models efficiently by involving three techniques:quantizing scheme search(QSS),quantizing precision learning(QPL),and quantized architecture generation(QAG).QSS introduces five quantizing schemes and defines three new schemes as a candidate set for scheme search,and then uses the Differentiable Neural Architecture Search(DNAS)algorithm to seek the layer-or model-desired scheme from the set.QPL is the first method to learn mixed-precision policies by reparameterizing the bitwidths of quantizing schemes,to the best of our knowledge.QPL optimizes both classification loss and precision loss of DNNs efficiently and obtains the relatively optimal mixed-precision model within limited model size and memory footprint.QAG is designed to convert arbitrary architectures into corresponding quantized ones without manual intervention,to facilitate end-to-end neural network quantization.We have implemented AutoQNN and integrated it into Keras.Extensive experiments demonstrate that AutoQNN can consistently outperform state-of-the-art quantization.For 2-bit weight and activation of AlexNet and ResNet18,AutoQNN can achieve the accuracy results of 59.75%and 68.86%,respectively,and obtain accuracy improvements by up to 1.65%and 1.74%,respectively,compared with state-of-the-art methods.Especially,compared with the full-precision AlexNet and ResNet18,the 2-bit models only slightly incur accuracy degradation by 0.26%and 0.76%,respectively,which can fulfill practical application demands.展开更多
Large-scale Language Models(LLMs)have achieved significant breakthroughs in Natural Language Processing(NLP),driven by the pre-training and fine-tuning paradigm.While this approach allows models to specialize in speci...Large-scale Language Models(LLMs)have achieved significant breakthroughs in Natural Language Processing(NLP),driven by the pre-training and fine-tuning paradigm.While this approach allows models to specialize in specific tasks with reduced training costs,the substantial memory requirements during fine-tuning present a barrier to broader deployment.Parameter-Efficient Fine-Tuning(PEFT)techniques,such as Low-Rank Adaptation(LoRA),and parameter quantization methods have emerged as solutions to address these challenges by optimizing memory usage and computational efficiency.Among these,QLoRA,which combines PEFT and quantization,has demonstrated notable success in reducing memory footprints during fine-tuning,prompting the development of various QLoRA variants.Despite these advancements,the quantitative impact of key variables on the fine-tuning performance of quantized LLMs remains underexplored.This study presents a comprehensive analysis of these key variables,focusing on their influence across different layer types and depths within LLM architectures.Our investigation uncovers several critical findings:(1)Larger layers,such as MLP layers,can maintain performance despite reductions in adapter rank,while smaller layers,like self-attention layers,aremore sensitive to such changes;(2)The effectiveness of balancing factors depends more on specific values rather than layer type or depth;(3)In quantization-aware fine-tuning,larger layers can effectively utilize smaller adapters,whereas smaller layers struggle to do so.These insights suggest that layer type is a more significant determinant of fine-tuning success than layer depth when optimizing quantized LLMs.Moreover,for the same discount of trainable parameters,reducing the trainable parameters in a larger layer is more effective in preserving fine-tuning accuracy than in a smaller one.This study provides valuable guidance for more efficient fine-tuning strategies and opens avenues for further research into optimizing LLM fine-tuning in resource-constrained environments.展开更多
Quantization noise caused by analog-to-digital converter(ADC)gives rise to the reliability performance degradation of communication systems.In this paper,a quantized non-Hermitian symmetry(NHS)orthogonal frequency-div...Quantization noise caused by analog-to-digital converter(ADC)gives rise to the reliability performance degradation of communication systems.In this paper,a quantized non-Hermitian symmetry(NHS)orthogonal frequency-division multiplexing-based visible light communication(OFDM-VLC)system is presented.In order to analyze the effect of the resolution of ADC on NHS OFDM-VLC,a quantized mathematical model of NHS OFDM-VLC is established.Based on the proposed quantized model,a closed-form bit error rate(BER)expression is derived.The theoretical analysis and simulation results both confirm the effectiveness of the obtained BER formula in high-resolution ADC.In addition,channel coding is helpful in compensating for the BER performance loss due to the utilization of lower resolution ADC.展开更多
The Tarim Basin has revealed numerous tight sandstone oil and gas reservoirs.The tidal fl at zone in the Shunbei area is currently in the detailed exploration stage,requiring a comprehensive description of the sand bo...The Tarim Basin has revealed numerous tight sandstone oil and gas reservoirs.The tidal fl at zone in the Shunbei area is currently in the detailed exploration stage,requiring a comprehensive description of the sand body distribution characteristics for rational exploration well deployment.However,using a single method for sand body prediction has yielded poor results.Seismic facies analysis can eff ectively predict the macro-development characteristics of sedimentary sand bodies but lacks the resolution to capture fine details.In contrast,single-well sedimentary facies analysis can describe detailed sand body development but struggles to reveal broader trends.Therefore,this study proposes a method that combines seismic facies analysis with single-well sedimentary microfacies analysis,using the lower section of the Kepingtage Formation in the Shunbei area as a case study.First,seismic facies were obtained through unsupervised vector quantization to control the macro-distribution characteristics of sand bodies,while principal component analysis(PCA)was applied to improve the depiction of fine sand body details from seismic attributes.Based on 3D seismic data,well-logging data,and geological interpretation results,a detailed structural interpretation was performed to establish a high-precision stratigraphic framework,thereby enhancing the accuracy of sand body prediction.Seismic facies analysis was then conducted to obtain the macro-distribution characteristics of the sand bodies.Subsequently,core data and logging curves from individual wells were used to clarify the vertical development characteristics of tidal channels and sandbars.Next,PCA was employed to select the seismic attributes most sensitive to sand bodies in diff erent sedimentary facies.Results indicate that RMS amplitude in the subtidal zone and instantaneous phase in the intertidal zone are the most sensitive to sand bodies.A comparative analysis of individual seismic attributes for sand body characterization revealed that facies-based delineation improved the accuracy of sand body identification,eff ectively capturing their contours and shapes.This method,which integrates seismic facies,single-well sedimentary microfacies,and machine learning techniques,enhances the precision of sand body characterization and off ers a novel approach to sand body prediction.展开更多
The rapid growth of digital data necessitates advanced natural language processing(NLP)models like BERT(Bidi-rectional Encoder Representations from Transformers),known for its superior performance in text classificati...The rapid growth of digital data necessitates advanced natural language processing(NLP)models like BERT(Bidi-rectional Encoder Representations from Transformers),known for its superior performance in text classification.However,BERT’s size and computational demands limit its practicality,especially in resource-constrained settings.This research compresses the BERT base model for Bengali emotion classification through knowledge distillation(KD),pruning,and quantization techniques.Despite Bengali being the sixth most spoken language globally,NLP research in this area is limited.Our approach addresses this gap by creating an efficient BERT-based model for Bengali text.We have explored 20 combinations for KD,quantization,and pruning,resulting in improved speedup,fewer parameters,and reduced memory size.Our best results demonstrate significant improvements in both speed and efficiency.For instance,in the case of mBERT,we achieved a 3.87×speedup and 4×compression ratio with a combination of Distil+Prune+Quant that reduced parameters from 178 to 46 M,while the memory size decreased from 711 to 178 MB.These results offer scalable solutions for NLP tasks in various languages and advance the field of model compression,making these models suitable for real-world applications in resource-limited environments.展开更多
A new phenomenological model(axionic QCD string)is constructed to study the topological issues of the QCD vacuum and hadron structure.It provides an alternative way of tackling the Strong CP problem,which is different...A new phenomenological model(axionic QCD string)is constructed to study the topological issues of the QCD vacuum and hadron structure.It provides an alternative way of tackling the Strong CP problem,which is different from the traditional Peccei–Quinn approach.Neither new particle nor extra symmetry is introduced,and the role of the Peccei–Quinn axion is played by a quasiparticle arising from the phase of the quark condensate,dubbed as axionic excitation.The derivative of this excitation field is decomposed into a regular part and a singular part,and the latter contains vorticity from the string configuration.A hidden gauge symmetry is revealed in this decomposition and vorticity is represented by an emergent gauge field associated with anomalies.These components,together with the anomaly-inflow mechanism,complete the effective Lagrangian description for the axionic QCD string.展开更多
Myocardial infarction(MI)is one of the leading causes of death globally among cardiovascular diseases,necessitating modern and accurate diagnostics for cardiac patient conditions.Among the available functional diagnos...Myocardial infarction(MI)is one of the leading causes of death globally among cardiovascular diseases,necessitating modern and accurate diagnostics for cardiac patient conditions.Among the available functional diagnostic methods,electrocardiography(ECG)is particularly well-known for its ability to detect MI.However,confirming its accuracy—particularly in identifying the localization of myocardial damage—often presents challenges in practice.This study,therefore,proposes a new approach based on machine learning models for the analysis of 12-lead ECG data to accurately identify the localization of MI.In particular,the learning vector quantization(LVQ)algorithm was applied,considering the contribution of each ECG lead in the 12-channel system,which obtained an accuracy of 87%in localizing damaged myocardium.The developed model was tested on verified data from the PTB database,including 445 ECG recordings from both healthy individuals and MI-diagnosed patients.The results demonstrated that the 12-lead ECG system allows for a comprehensive understanding of cardiac activities in myocardial infarction patients,serving as an essential tool for the diagnosis of myocardial conditions and localizing their damage.A comprehensive comparison was performed,including CNN,SVM,and Logistic Regression,to evaluate the proposed LVQ model.The results demonstrate that the LVQ model achieves competitive performance in diagnostic tasks while maintaining computational efficiency,making it suitable for resource-constrained environments.This study also applies a carefully designed data pre-processing flow,including class balancing and noise removal,which improves the reliability and reproducibility of the results.These aspects highlight the potential application of the LVQ model in cardiac diagnostics,opening up prospects for its use along with more complex neural network architectures.展开更多
We consider a relativistic two-fluid model of superfluidity,in which the superfluid is described by an order parameter that is a complex scalar field satisfying the nonlinear Klein-Gordon equation(NLKG).The coupling t...We consider a relativistic two-fluid model of superfluidity,in which the superfluid is described by an order parameter that is a complex scalar field satisfying the nonlinear Klein-Gordon equation(NLKG).The coupling to the normal fluid is introduced via a covariant current-current interaction,which results in the addition of an effective potential,whose imaginary part describes particle transfer between superfluid and normal fluid.Quantized vorticity arises in a class of singular solutions and the related vortex dynamics is incorporated in the modified NLKG,facilitating numerical analysis which is usually very complicated in the phenomenology of vortex filaments.The dual transformation to a string theory description(Kalb-Ramond)of quantum vorticity,the Magnus force,and the mutual friction between quantized vortices and normal fluid are also studied.展开更多
This paper proposes a novel method for the automatic diagnosis of keratitis using feature vector quantization and self-attention mechanisms(ADK_FVQSAM).First,high-level features are extracted using the DenseNet121 bac...This paper proposes a novel method for the automatic diagnosis of keratitis using feature vector quantization and self-attention mechanisms(ADK_FVQSAM).First,high-level features are extracted using the DenseNet121 backbone network,followed by adaptive average pooling to scale the features to a fixed length.Subsequently,product quantization with residuals(PQR)is applied to convert continuous feature vectors into discrete features representations,preserving essential information insensitive to image quality variations.The quantized and original features are concatenated and fed into a self-attention mechanism to capture keratitis-related features.Finally,these enhanced features are classified through a fully connected layer.Experiments on clinical low-quality(LQ)images show that ADK_FVQSAM achieves accuracies of 87.7%,81.9%,and 89.3% for keratitis,other corneal abnormalities,and normal corneas,respectively.Compared to DenseNet121,Swin transformer,and InceptionResNet,ADK_FVQSAM improves average accuracy by 3.1%,11.3%,and 15.3%,respectively.These results demonstrate that ADK_FVQSAM significantly enhances the recognition performance of keratitis based on LQ slit-lamp images,offering a practical approach for clinical application.展开更多
In this paper,a distributed Event-Triggered(ET)collision avoidance coordinated control for Quadrotor Unmanned Aerial Vehicles(QUAVs)is proposed based on Virtual Tubes(VTs)with flexible boundaries in the presence of un...In this paper,a distributed Event-Triggered(ET)collision avoidance coordinated control for Quadrotor Unmanned Aerial Vehicles(QUAVs)is proposed based on Virtual Tubes(VTs)with flexible boundaries in the presence of unknown external disturbances.Firstly,VTs are constructed for each QUAV,and the QUAV is restricted into the corresponding VT by the artificial potential field,which is distributed around the boundary of the VT.Thus,the collisions between QUAVs are avoided.Besides,the boundaries of the VTs are flexible by the modification signals,which are generated by the self-regulating auxiliary systems,to make the repulsive force smaller and give more buffer space for QUAVs without collision.Then,a novel ET mechanism is designed by introducing the concept of prediction to the traditional fixed threshold ET mechanism.Furthermore,a disturbance observer is proposed to deal with the adverse effects of the unknown external disturbance.On this basis,a distributed ET collision avoidance coordinated controller is proposed.Then,the proposed controller is quantized by the hysteresis uniform quantizer and then sent to the actuator only at the ET instants.The boundedness of the closed-loop signals is verified by the Lyapunov method.Finally,simulation and experimental results are performed to demonstrate the superiority of the proposed control method.展开更多
This paper is concerned with distributed Nash equi librium seeking strategies under quantized communication. In the proposed seeking strategy, a projection operator is synthesized with a gradient search method to achi...This paper is concerned with distributed Nash equi librium seeking strategies under quantized communication. In the proposed seeking strategy, a projection operator is synthesized with a gradient search method to achieve the optimization o players' objective functions while restricting their actions within required non-empty, convex and compact domains. In addition, a leader-following consensus protocol, in which quantized informa tion flows are utilized, is employed for information sharing among players. More specifically, logarithmic quantizers and uniform quantizers are investigated under both undirected and connected communication graphs and strongly connected digraphs, respec tively. Through Lyapunov stability analysis, it is shown that play ers' actions can be steered to a neighborhood of the Nash equilib rium with logarithmic and uniform quantizers, and the quanti fied convergence error depends on the parameter of the quan tizer for both undirected and directed cases. A numerical exam ple is given to verify the theoretical results.展开更多
Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,maki...Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,making it a widely adopted approach in various industrial fields.This paper mainly studied the defect detection method for nonwoven materials based on the improved Nano Det-Plus model.Using the constructed samples of defects in nonwoven materials as the research objects,transfer learning experiments were conducted based on the Nano DetPlus object detection framework.Within this framework,the Backbone,path aggregation feature pyramid network(PAFPN)and Head network models were compared and trained through a process of freezing,with the ultimate aim of bolstering the model's feature extraction abilities and elevating detection accuracy.The half-precision quantization method was used to optimize the model after transfer learning experiments,reducing model weights and computational complexity to improve the detection speed.Performance comparisons were conducted between the improved model and the original Nano Det-Plus model,YOLO,SSD and other common industrial defect detection algorithms,validating that the improved methods based on transfer learning and semi-precision quantization enabled the model to meet the practical requirements of industrial production.展开更多
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 embedd...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.展开更多
Federated Learning(FL)is an emerging machine learning framework designed to preserve privacy.However,the continuous updating of model parameters over uplink channels with limited throughput leads to a huge communicati...Federated Learning(FL)is an emerging machine learning framework designed to preserve privacy.However,the continuous updating of model parameters over uplink channels with limited throughput leads to a huge communication overload,which is a major challenge for FL.To address this issue,we propose an adaptive gradient quantization approach that enhances communication efficiency.Aiming to minimize the total communication costs,we consider both the correlation of gradients between local clients and the correlation of gradients between communication rounds,namely,in the time and space dimensions.The compression strategy is based on rate distortion theory,which allows us to find an optimal quantization strategy for the gradients.To further reduce the computational complexity,we introduce the Kalman filter into the proposed approach.Finally,numerical results demonstrate the effectiveness and robustness of the proposed rate-distortion optimization adaptive gradient quantization approach in significantly reducing the communication costs when compared to other quantization methods.展开更多
In this tutorial paper, we explore the field of quantized feedback control, which has gained significant attention due to the growing prevalence of networked control systems. These systems require the transmission of ...In this tutorial paper, we explore the field of quantized feedback control, which has gained significant attention due to the growing prevalence of networked control systems. These systems require the transmission of feedback information, such as measurements and control signals, over digital networks, presenting novel challenges in estimation and control design. Our examination encompasses various topics, including the minimal information needed for effective feedback control, the design of quantizers, strategies for quantized control design and estimation,achieving consensus control with quantized data, and the pursuit of high-precision tracking using quantized measurements.展开更多
We propose a joint look-up-table(LUT)-based nonlinear predistortion and digital resolution enhancement scheme to achieve high-speed and low-cost optical interconnects using low-resolution digital-to-analog converters(...We propose a joint look-up-table(LUT)-based nonlinear predistortion and digital resolution enhancement scheme to achieve high-speed and low-cost optical interconnects using low-resolution digital-to-analog converters(DACs).The LUT-based predistortion is employed to mitigate the patterndependent effect(PDE)of a semiconductor optical amplifier(SOA),while the digital resolution enhancer(DRE)is utilized to shape the quantization noise,lowering the requirement for the resolution of DAC.We experimentally demonstrate O-band intensity modulation and direct detection(IM/DD)transmission of 124-GBd 4∕6-level pulse-amplitude modulation ePAMT-4∕6 and 112-GBd PAM-8 signals over a 2-km standard single-mode fiber(SSMF)with 3∕3.5∕4-bit DACs.In the case of 40-km SSMF transmission with an SOAbased preamplifier,124-GBd on-off-keying(OOK)/PAM-3/PAM-4 signals are successfully transmitted with 1.5∕2∕3-bit DACs.To the best of our knowledge,we have achieved the highest net data rates of 235.3-Gb∕s PAM-4,289.7-Gb∕s PAM-6,and 294.7 Gb∕s PAM-8 signals over 2-km SSMF,as well as 117.6-Gb∕s OOK,173.8-Gb∕s PAM-3,and−231.8 Gb∕s PAM-4 signals over 40-km SSMF,employing low-resolution DACs.The experimental results reveal that the joint LUT-based predistortion and DRE effectively mitigate the PDE and improve the signal-to-quantization noise ratio by shaping the noise.The proposed scheme can provide a powerful solution for low-cost IM/DD optical interconnects beyond 200 Gb∕s.展开更多
In networked system identification,how to effectively use communication resources and improve convergence speed is the focus of attention.However,there is an inherent contradiction between the two tasks.In this paper,...In networked system identification,how to effectively use communication resources and improve convergence speed is the focus of attention.However,there is an inherent contradiction between the two tasks.In this paper,the event-driven communication is used to save communication resources for the identification of finite impulse response systems,and the input design is carried out to meet the requirements of convergence speed.First,a difference-driven communication is proposed.Then,the performance of the communication mechanism is analyzed,and the calculation method of its communication rate is given.After that,according to the communication rate and the convergence rate of the identification algorithm,the input design problem is transformed into a constrained optimization problem,and the algorithm for finding the optimal solution is given.In addition,considering the case that the output is quantized by multiple thresholds,the way to calculate its communication rate is given and the influence of threshold number on communication rate is discussed.Finally,the effectiveness of the algorithm is verified by simulation.展开更多
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 re...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.展开更多
文摘On the basis of realization of beach information and its differentiating of high-resolution remote sensing image on coastal zone, extracting objects are carried through RS multi-scale diagnostic analysis, and fast information extraction methods and key technologies are put forward. Meanwhile image segmentation methods are set forth for objects of coastal zone. And through the application of Otsu2D to the segmentation of water area and dock and the applying of Gabor filter to the separation and extraction of construction, some typical applications of high-resolution RS image are presented in the field of coastal zone surface objects＇ recognition. Quantizing high-resolution RS information on the coastal zone proved to be of great scientific and practical significance for coastal development and management.
文摘A rule of how the ignition probability or invalidation probability of infantry grenade varies with the projectile quantity is presented based on the statistical analysis. The statistical induction, quantizing analysis and data processing in the infantry grenade type approval test are completed, and we obtain how various test factors in the research item affect the test results(invalidation probability), i.e., quantized data. The acquisition of these quantized data provides theory basis and data auspice for further reasonable filtration of all test factors and the optimization of test scheme.
基金supported by the China Postdoctoral Science Foundation under Grant No.2022M721707the National Natural Science Foundation of China under Grant Nos.62002175 and 62272248+1 种基金the Special Funding for Excellent Enterprise Technology Correspondent of Tianjin under Grant No.21YDTPJC00380the Open Project Foundation of Information Security Evaluation Center of Civil Aviation,Civil Aviation University of China,under Grant No.ISECCA-202102.
文摘Exploring the expected quantizing scheme with suitable mixed-precision policy is the key to compress deep neural networks(DNNs)in high efficiency and accuracy.This exploration implies heavy workloads for domain experts,and an automatic compression method is needed.However,the huge search space of the automatic method introduces plenty of computing budgets that make the automatic process challenging to be applied in real scenarios.In this paper,we propose an end-to-end framework named AutoQNN,for automatically quantizing different layers utilizing different schemes and bitwidths without any human labor.AutoQNN can seek desirable quantizing schemes and mixed-precision policies for mainstream DNN models efficiently by involving three techniques:quantizing scheme search(QSS),quantizing precision learning(QPL),and quantized architecture generation(QAG).QSS introduces five quantizing schemes and defines three new schemes as a candidate set for scheme search,and then uses the Differentiable Neural Architecture Search(DNAS)algorithm to seek the layer-or model-desired scheme from the set.QPL is the first method to learn mixed-precision policies by reparameterizing the bitwidths of quantizing schemes,to the best of our knowledge.QPL optimizes both classification loss and precision loss of DNNs efficiently and obtains the relatively optimal mixed-precision model within limited model size and memory footprint.QAG is designed to convert arbitrary architectures into corresponding quantized ones without manual intervention,to facilitate end-to-end neural network quantization.We have implemented AutoQNN and integrated it into Keras.Extensive experiments demonstrate that AutoQNN can consistently outperform state-of-the-art quantization.For 2-bit weight and activation of AlexNet and ResNet18,AutoQNN can achieve the accuracy results of 59.75%and 68.86%,respectively,and obtain accuracy improvements by up to 1.65%and 1.74%,respectively,compared with state-of-the-art methods.Especially,compared with the full-precision AlexNet and ResNet18,the 2-bit models only slightly incur accuracy degradation by 0.26%and 0.76%,respectively,which can fulfill practical application demands.
基金supported by the National Key R&D Program of China(No.2021YFB0301200)National Natural Science Foundation of China(No.62025208).
文摘Large-scale Language Models(LLMs)have achieved significant breakthroughs in Natural Language Processing(NLP),driven by the pre-training and fine-tuning paradigm.While this approach allows models to specialize in specific tasks with reduced training costs,the substantial memory requirements during fine-tuning present a barrier to broader deployment.Parameter-Efficient Fine-Tuning(PEFT)techniques,such as Low-Rank Adaptation(LoRA),and parameter quantization methods have emerged as solutions to address these challenges by optimizing memory usage and computational efficiency.Among these,QLoRA,which combines PEFT and quantization,has demonstrated notable success in reducing memory footprints during fine-tuning,prompting the development of various QLoRA variants.Despite these advancements,the quantitative impact of key variables on the fine-tuning performance of quantized LLMs remains underexplored.This study presents a comprehensive analysis of these key variables,focusing on their influence across different layer types and depths within LLM architectures.Our investigation uncovers several critical findings:(1)Larger layers,such as MLP layers,can maintain performance despite reductions in adapter rank,while smaller layers,like self-attention layers,aremore sensitive to such changes;(2)The effectiveness of balancing factors depends more on specific values rather than layer type or depth;(3)In quantization-aware fine-tuning,larger layers can effectively utilize smaller adapters,whereas smaller layers struggle to do so.These insights suggest that layer type is a more significant determinant of fine-tuning success than layer depth when optimizing quantized LLMs.Moreover,for the same discount of trainable parameters,reducing the trainable parameters in a larger layer is more effective in preserving fine-tuning accuracy than in a smaller one.This study provides valuable guidance for more efficient fine-tuning strategies and opens avenues for further research into optimizing LLM fine-tuning in resource-constrained environments.
基金supported by the National Natural Science Foundation of China(No.62201508)the Zhejiang Provincial Natural Science Foundation of China(Nos.LZ21F010001 and LQ23F010004)the State Key Laboratory of Millimeter Waves of Southeast University,China(No.K202212).
文摘Quantization noise caused by analog-to-digital converter(ADC)gives rise to the reliability performance degradation of communication systems.In this paper,a quantized non-Hermitian symmetry(NHS)orthogonal frequency-division multiplexing-based visible light communication(OFDM-VLC)system is presented.In order to analyze the effect of the resolution of ADC on NHS OFDM-VLC,a quantized mathematical model of NHS OFDM-VLC is established.Based on the proposed quantized model,a closed-form bit error rate(BER)expression is derived.The theoretical analysis and simulation results both confirm the effectiveness of the obtained BER formula in high-resolution ADC.In addition,channel coding is helpful in compensating for the BER performance loss due to the utilization of lower resolution ADC.
基金Collaborative Project Grant from the Exploration and Development Research Institute of SINOPEC Northwest Oilfi eld Company(Grant No.KY2021-S-104).
文摘The Tarim Basin has revealed numerous tight sandstone oil and gas reservoirs.The tidal fl at zone in the Shunbei area is currently in the detailed exploration stage,requiring a comprehensive description of the sand body distribution characteristics for rational exploration well deployment.However,using a single method for sand body prediction has yielded poor results.Seismic facies analysis can eff ectively predict the macro-development characteristics of sedimentary sand bodies but lacks the resolution to capture fine details.In contrast,single-well sedimentary facies analysis can describe detailed sand body development but struggles to reveal broader trends.Therefore,this study proposes a method that combines seismic facies analysis with single-well sedimentary microfacies analysis,using the lower section of the Kepingtage Formation in the Shunbei area as a case study.First,seismic facies were obtained through unsupervised vector quantization to control the macro-distribution characteristics of sand bodies,while principal component analysis(PCA)was applied to improve the depiction of fine sand body details from seismic attributes.Based on 3D seismic data,well-logging data,and geological interpretation results,a detailed structural interpretation was performed to establish a high-precision stratigraphic framework,thereby enhancing the accuracy of sand body prediction.Seismic facies analysis was then conducted to obtain the macro-distribution characteristics of the sand bodies.Subsequently,core data and logging curves from individual wells were used to clarify the vertical development characteristics of tidal channels and sandbars.Next,PCA was employed to select the seismic attributes most sensitive to sand bodies in diff erent sedimentary facies.Results indicate that RMS amplitude in the subtidal zone and instantaneous phase in the intertidal zone are the most sensitive to sand bodies.A comparative analysis of individual seismic attributes for sand body characterization revealed that facies-based delineation improved the accuracy of sand body identification,eff ectively capturing their contours and shapes.This method,which integrates seismic facies,single-well sedimentary microfacies,and machine learning techniques,enhances the precision of sand body characterization and off ers a novel approach to sand body prediction.
文摘The rapid growth of digital data necessitates advanced natural language processing(NLP)models like BERT(Bidi-rectional Encoder Representations from Transformers),known for its superior performance in text classification.However,BERT’s size and computational demands limit its practicality,especially in resource-constrained settings.This research compresses the BERT base model for Bengali emotion classification through knowledge distillation(KD),pruning,and quantization techniques.Despite Bengali being the sixth most spoken language globally,NLP research in this area is limited.Our approach addresses this gap by creating an efficient BERT-based model for Bengali text.We have explored 20 combinations for KD,quantization,and pruning,resulting in improved speedup,fewer parameters,and reduced memory size.Our best results demonstrate significant improvements in both speed and efficiency.For instance,in the case of mBERT,we achieved a 3.87×speedup and 4×compression ratio with a combination of Distil+Prune+Quant that reduced parameters from 178 to 46 M,while the memory size decreased from 711 to 178 MB.These results offer scalable solutions for NLP tasks in various languages and advance the field of model compression,making these models suitable for real-world applications in resource-limited environments.
基金supported by the Natural Science Foundation of Fujian Province(Grant No.2022J011130)the Research Starting Grant from Minjiang University(Grant No.30804317)。
文摘A new phenomenological model(axionic QCD string)is constructed to study the topological issues of the QCD vacuum and hadron structure.It provides an alternative way of tackling the Strong CP problem,which is different from the traditional Peccei–Quinn approach.Neither new particle nor extra symmetry is introduced,and the role of the Peccei–Quinn axion is played by a quasiparticle arising from the phase of the quark condensate,dubbed as axionic excitation.The derivative of this excitation field is decomposed into a regular part and a singular part,and the latter contains vorticity from the string configuration.A hidden gauge symmetry is revealed in this decomposition and vorticity is represented by an emergent gauge field associated with anomalies.These components,together with the anomaly-inflow mechanism,complete the effective Lagrangian description for the axionic QCD string.
基金funded by the Ministry of Science and Higher Education of the Republic of Kazakhstan,grant numbers AP14969403 and AP23485820.
文摘Myocardial infarction(MI)is one of the leading causes of death globally among cardiovascular diseases,necessitating modern and accurate diagnostics for cardiac patient conditions.Among the available functional diagnostic methods,electrocardiography(ECG)is particularly well-known for its ability to detect MI.However,confirming its accuracy—particularly in identifying the localization of myocardial damage—often presents challenges in practice.This study,therefore,proposes a new approach based on machine learning models for the analysis of 12-lead ECG data to accurately identify the localization of MI.In particular,the learning vector quantization(LVQ)algorithm was applied,considering the contribution of each ECG lead in the 12-channel system,which obtained an accuracy of 87%in localizing damaged myocardium.The developed model was tested on verified data from the PTB database,including 445 ECG recordings from both healthy individuals and MI-diagnosed patients.The results demonstrated that the 12-lead ECG system allows for a comprehensive understanding of cardiac activities in myocardial infarction patients,serving as an essential tool for the diagnosis of myocardial conditions and localizing their damage.A comprehensive comparison was performed,including CNN,SVM,and Logistic Regression,to evaluate the proposed LVQ model.The results demonstrate that the LVQ model achieves competitive performance in diagnostic tasks while maintaining computational efficiency,making it suitable for resource-constrained environments.This study also applies a carefully designed data pre-processing flow,including class balancing and noise removal,which improves the reliability and reproducibility of the results.These aspects highlight the potential application of the LVQ model in cardiac diagnostics,opening up prospects for its use along with more complex neural network architectures.
文摘We consider a relativistic two-fluid model of superfluidity,in which the superfluid is described by an order parameter that is a complex scalar field satisfying the nonlinear Klein-Gordon equation(NLKG).The coupling to the normal fluid is introduced via a covariant current-current interaction,which results in the addition of an effective potential,whose imaginary part describes particle transfer between superfluid and normal fluid.Quantized vorticity arises in a class of singular solutions and the related vortex dynamics is incorporated in the modified NLKG,facilitating numerical analysis which is usually very complicated in the phenomenology of vortex filaments.The dual transformation to a string theory description(Kalb-Ramond)of quantum vorticity,the Magnus force,and the mutual friction between quantized vortices and normal fluid are also studied.
基金supported by the National Natural Science Foundation of China(Nos.62276210,82201148 and 62376215)the Key Research and Development Project of Shaanxi Province(No.2025CY-YBXM-044)+3 种基金the Natural Science Foundation of Zhejiang Province(No.LQ22H120002)the Medical Health Science and Technology Project of Zhejiang Province(Nos.2022RC069 and 2023KY1140)the Natural Science Foundation of Ningbo(No.2023J390)the Ningbo Top Medical and Health Research Program(No.2023030716).
文摘This paper proposes a novel method for the automatic diagnosis of keratitis using feature vector quantization and self-attention mechanisms(ADK_FVQSAM).First,high-level features are extracted using the DenseNet121 backbone network,followed by adaptive average pooling to scale the features to a fixed length.Subsequently,product quantization with residuals(PQR)is applied to convert continuous feature vectors into discrete features representations,preserving essential information insensitive to image quality variations.The quantized and original features are concatenated and fed into a self-attention mechanism to capture keratitis-related features.Finally,these enhanced features are classified through a fully connected layer.Experiments on clinical low-quality(LQ)images show that ADK_FVQSAM achieves accuracies of 87.7%,81.9%,and 89.3% for keratitis,other corneal abnormalities,and normal corneas,respectively.Compared to DenseNet121,Swin transformer,and InceptionResNet,ADK_FVQSAM improves average accuracy by 3.1%,11.3%,and 15.3%,respectively.These results demonstrate that ADK_FVQSAM significantly enhances the recognition performance of keratitis based on LQ slit-lamp images,offering a practical approach for clinical application.
基金supported in part by the National Key R&D Program of China(No.2023YFB4704400)in part by the National Natural Science Foundation of China(Nos.U23B2036,U2013201).
文摘In this paper,a distributed Event-Triggered(ET)collision avoidance coordinated control for Quadrotor Unmanned Aerial Vehicles(QUAVs)is proposed based on Virtual Tubes(VTs)with flexible boundaries in the presence of unknown external disturbances.Firstly,VTs are constructed for each QUAV,and the QUAV is restricted into the corresponding VT by the artificial potential field,which is distributed around the boundary of the VT.Thus,the collisions between QUAVs are avoided.Besides,the boundaries of the VTs are flexible by the modification signals,which are generated by the self-regulating auxiliary systems,to make the repulsive force smaller and give more buffer space for QUAVs without collision.Then,a novel ET mechanism is designed by introducing the concept of prediction to the traditional fixed threshold ET mechanism.Furthermore,a disturbance observer is proposed to deal with the adverse effects of the unknown external disturbance.On this basis,a distributed ET collision avoidance coordinated controller is proposed.Then,the proposed controller is quantized by the hysteresis uniform quantizer and then sent to the actuator only at the ET instants.The boundedness of the closed-loop signals is verified by the Lyapunov method.Finally,simulation and experimental results are performed to demonstrate the superiority of the proposed control method.
基金supported by the National Natural Science Foundation of China (NSFC)(62222308, 62173181, 62073171, 62221004)the Natural Science Foundation of Jiangsu Province (BK20200744, BK20220139)+3 种基金Jiangsu Specially-Appointed Professor (RK043STP19001)the Young Elite Scientists Sponsorship Program by CAST (2021QNRC001)1311 Talent Plan of Nanjing University of Posts and Telecommunicationsthe Fundamental Research Funds for the Central Universities (30920032203)。
文摘This paper is concerned with distributed Nash equi librium seeking strategies under quantized communication. In the proposed seeking strategy, a projection operator is synthesized with a gradient search method to achieve the optimization o players' objective functions while restricting their actions within required non-empty, convex and compact domains. In addition, a leader-following consensus protocol, in which quantized informa tion flows are utilized, is employed for information sharing among players. More specifically, logarithmic quantizers and uniform quantizers are investigated under both undirected and connected communication graphs and strongly connected digraphs, respec tively. Through Lyapunov stability analysis, it is shown that play ers' actions can be steered to a neighborhood of the Nash equilib rium with logarithmic and uniform quantizers, and the quanti fied convergence error depends on the parameter of the quan tizer for both undirected and directed cases. A numerical exam ple is given to verify the theoretical results.
基金National Key Research and Development Program of China(Nos.2022YFB4700600 and 2022YFB4700605)National Natural Science Foundation of China(Nos.61771123 and 62171116)+1 种基金Fundamental Research Funds for the Central UniversitiesGraduate Student Innovation Fund of Donghua University,China(No.CUSF-DH-D-2022044)。
文摘Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,making it a widely adopted approach in various industrial fields.This paper mainly studied the defect detection method for nonwoven materials based on the improved Nano Det-Plus model.Using the constructed samples of defects in nonwoven materials as the research objects,transfer learning experiments were conducted based on the Nano DetPlus object detection framework.Within this framework,the Backbone,path aggregation feature pyramid network(PAFPN)and Head network models were compared and trained through a process of freezing,with the ultimate aim of bolstering the model's feature extraction abilities and elevating detection accuracy.The half-precision quantization method was used to optimize the model after transfer learning experiments,reducing model weights and computational complexity to improve the detection speed.Performance comparisons were conducted between the improved model and the original Nano Det-Plus model,YOLO,SSD and other common industrial defect detection algorithms,validating that the improved methods based on transfer learning and semi-precision quantization enabled the model to meet the practical requirements of industrial production.
基金The Qian Xuesen Youth Innovation Foundation from China Aerospace Science and Technology Corporation(Grant Number 2022JY51).
文摘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.
基金supported in part by the Key Research and Development Program of Jiangsu Province(Grant No.BE2020084-2)in part by the National Key Research and Development Program of China(Grant No.2020YFB1600104)+4 种基金in part by the Key Research and Development Special Project of school and local cooperation in Lvliang(Grant No.2023XDHZ18)in part by Southeast University-China Mobile Research Institute Joint Innovation Centerin part by the National Natural Science Foundation of China(Grant No.62371119)in part by the Key Research and Development Program of Jiangsu Province(Grant No.BE2022059-3)in part by the Zhi Shan Young Scholar Program of Southeast University。
文摘Federated Learning(FL)is an emerging machine learning framework designed to preserve privacy.However,the continuous updating of model parameters over uplink channels with limited throughput leads to a huge communication overload,which is a major challenge for FL.To address this issue,we propose an adaptive gradient quantization approach that enhances communication efficiency.Aiming to minimize the total communication costs,we consider both the correlation of gradients between local clients and the correlation of gradients between communication rounds,namely,in the time and space dimensions.The compression strategy is based on rate distortion theory,which allows us to find an optimal quantization strategy for the gradients.To further reduce the computational complexity,we introduce the Kalman filter into the proposed approach.Finally,numerical results demonstrate the effectiveness and robustness of the proposed rate-distortion optimization adaptive gradient quantization approach in significantly reducing the communication costs when compared to other quantization methods.
基金partially supported by National Natura Science Foundation of China (62350710214, U23A20325)Shenzhen Key Laboratory of Control Theory and Intelligent Systems (ZDSYS20220330161800001)。
文摘In this tutorial paper, we explore the field of quantized feedback control, which has gained significant attention due to the growing prevalence of networked control systems. These systems require the transmission of feedback information, such as measurements and control signals, over digital networks, presenting novel challenges in estimation and control design. Our examination encompasses various topics, including the minimal information needed for effective feedback control, the design of quantizers, strategies for quantized control design and estimation,achieving consensus control with quantized data, and the pursuit of high-precision tracking using quantized measurements.
基金supported by the National Key R&D Program of China(Grant No.2020YFB1806400)the China Scholarship Council(Grant No.202306230183)the National Natural Science Foundation of China(Grant Nos.62271305 and 62001287).
文摘We propose a joint look-up-table(LUT)-based nonlinear predistortion and digital resolution enhancement scheme to achieve high-speed and low-cost optical interconnects using low-resolution digital-to-analog converters(DACs).The LUT-based predistortion is employed to mitigate the patterndependent effect(PDE)of a semiconductor optical amplifier(SOA),while the digital resolution enhancer(DRE)is utilized to shape the quantization noise,lowering the requirement for the resolution of DAC.We experimentally demonstrate O-band intensity modulation and direct detection(IM/DD)transmission of 124-GBd 4∕6-level pulse-amplitude modulation ePAMT-4∕6 and 112-GBd PAM-8 signals over a 2-km standard single-mode fiber(SSMF)with 3∕3.5∕4-bit DACs.In the case of 40-km SSMF transmission with an SOAbased preamplifier,124-GBd on-off-keying(OOK)/PAM-3/PAM-4 signals are successfully transmitted with 1.5∕2∕3-bit DACs.To the best of our knowledge,we have achieved the highest net data rates of 235.3-Gb∕s PAM-4,289.7-Gb∕s PAM-6,and 294.7 Gb∕s PAM-8 signals over 2-km SSMF,as well as 117.6-Gb∕s OOK,173.8-Gb∕s PAM-3,and−231.8 Gb∕s PAM-4 signals over 40-km SSMF,employing low-resolution DACs.The experimental results reveal that the joint LUT-based predistortion and DRE effectively mitigate the PDE and improve the signal-to-quantization noise ratio by shaping the noise.The proposed scheme can provide a powerful solution for low-cost IM/DD optical interconnects beyond 200 Gb∕s.
基金supported in part by the National Natural Science Foundation of China(No.62173030)in part by the Beijing Natural Science Foundation(No.4222050).
文摘In networked system identification,how to effectively use communication resources and improve convergence speed is the focus of attention.However,there is an inherent contradiction between the two tasks.In this paper,the event-driven communication is used to save communication resources for the identification of finite impulse response systems,and the input design is carried out to meet the requirements of convergence speed.First,a difference-driven communication is proposed.Then,the performance of the communication mechanism is analyzed,and the calculation method of its communication rate is given.After that,according to the communication rate and the convergence rate of the identification algorithm,the input design problem is transformed into a constrained optimization problem,and the algorithm for finding the optimal solution is given.In addition,considering the case that the output is quantized by multiple thresholds,the way to calculate its communication rate is given and the influence of threshold number on communication rate is discussed.Finally,the effectiveness of the algorithm is verified by simulation.
基金Project supported by the National Natural Science Foundation of China(Grant Nos.12104414,12122412,12104464,and 12104413)the China Postdoctoral Science Foundation(Grant No.2021M702955).
文摘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.
