Synaptic pruning is a crucial process in synaptic refinement,eliminating unstable synaptic connections in neural circuits.This process is triggered and regulated primarily by spontaneous neural activity and experience...Synaptic pruning is a crucial process in synaptic refinement,eliminating unstable synaptic connections in neural circuits.This process is triggered and regulated primarily by spontaneous neural activity and experience-dependent mechanisms.The pruning process involves multiple molecular signals and a series of regulatory activities governing the“eat me”and“don't eat me”states.Under physiological conditions,the interaction between glial cells and neurons results in the clearance of unnecessary synapses,maintaining normal neural circuit functionality via synaptic pruning.Alterations in genetic and environmental factors can lead to imbalanced synaptic pruning,thus promoting the occurrence and development of autism spectrum disorder,schizophrenia,Alzheimer's disease,and other neurological disorders.In this review,we investigated the molecular mechanisms responsible for synaptic pruning during neural development.We focus on how synaptic pruning can regulate neural circuits and its association with neurological disorders.Furthermore,we discuss the application of emerging optical and imaging technologies to observe synaptic structure and function,as well as their potential for clinical translation.Our aim was to enhance our understanding of synaptic pruning during neural development,including the molecular basis underlying the regulation of synaptic function and the dynamic changes in synaptic density,and to investigate the potential role of these mechanisms in the pathophysiology of neurological diseases,thus providing a theoretical foundation for the treatment of neurological disorders.展开更多
Activation pruning reduces neural network complexity by eliminating low-importance neuron activations,yet identifying the critical pruning threshold—beyond which accuracy rapidly deteriorates—remains computationally...Activation pruning reduces neural network complexity by eliminating low-importance neuron activations,yet identifying the critical pruning threshold—beyond which accuracy rapidly deteriorates—remains computationally expensive and typically requires exhaustive search.We introduce a thermodynamics-inspired framework that treats activation distributions as energy-filtered physical systems and employs the free energy of activations as a principled evaluation metric.Phase-transition-like phenomena in the free-energy profile—such as extrema,inflection points,and curvature changes—yield reliable estimates of the critical pruning threshold,providing a theoretically grounded means of predicting sharp accuracy degradation.To further enhance efficiency,we propose a renormalized free energy technique that approximates full-evaluation free energy using only the activation distribution of the unpruned network.This eliminates repeated forward passes,dramatically reducing computational overhead and achieving speedups of up to 550×for MLPs.Extensive experiments across diverse vision architectures(MLP,CNN,ResNet,MobileNet,Vision Transformer)and text models(LSTM,BERT,ELECTRA,T5,GPT-2)on multiple datasets validate the generality,robustness,and computational efficiency of our approach.Overall,this work establishes a theoretically grounded and practically effective framework for activation pruning,bridging the gap between analytical understanding and efficient deployment of sparse neural networks.展开更多
Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subn...Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subnetworks in order to mitigate the exposure of sensitive data and reduce the overhead on client devices,thereby making SL particularly suitable for resource-constrained devices.Although SL prevents the direct transmission of raw data,it does not alleviate entirely the risk of privacy breaches.In fact,the data intermediately transmitted to the server sub-model may include patterns or information that could reveal sensitive data.Moreover,achieving a balance between model utility and data privacy has emerged as a challenging problem.In this article,we propose a novel defense approach that combines:(i)Adversarial learning,and(ii)Network channel pruning.In particular,the proposed adversarial learning approach is specifically designed to reduce the risk of private data exposure while maintaining high performance for the utility task.On the other hand,the suggested channel pruning enables the model to adaptively adjust and reactivate pruned channels while conducting adversarial training.The integration of these two techniques reduces the informativeness of the intermediate data transmitted by the client sub-model,thereby enhancing its robustness against attribute inference attacks without adding significant computational overhead,making it wellsuited for IoT devices,mobile platforms,and Internet of Vehicles(IoV)scenarios.The proposed defense approach was evaluated using EfficientNet-B0,a widely adopted compact model,along with three benchmark datasets.The obtained results showcased its superior defense capability against attribute inference attacks compared to existing state-of-the-art methods.This research’s findings demonstrated the effectiveness of the proposed channel pruning-based adversarial training approach in achieving the intended compromise between utility and privacy within SL frameworks.In fact,the classification accuracy attained by the attackers witnessed a drastic decrease of 70%.展开更多
The surge of large-scale models in recent years has led to breakthroughs in numerous fields,but it has also introduced higher computational costs and more complex network architectures.These increasingly large and int...The surge of large-scale models in recent years has led to breakthroughs in numerous fields,but it has also introduced higher computational costs and more complex network architectures.These increasingly large and intricate networks pose challenges for deployment and execution while also exacerbating the issue of network over-parameterization.To address this issue,various network compression techniques have been developed,such as network pruning.A typical pruning algorithm follows a three-step pipeline involving training,pruning,and retraining.Existing methods often directly set the pruned filters to zero during retraining,significantly reducing the parameter space.However,this direct pruning strategy frequently results in irreversible information loss.In the early stages of training,a network still contains much uncertainty,and evaluating filter importance may not be sufficiently rigorous.To manage the pruning process effectively,this paper proposes a flexible neural network pruning algorithm based on the logistic growth differential equation,considering the characteristics of network training.Unlike other pruning algorithms that directly reduce filter weights,this algorithm introduces a three-stage adaptive weight decay strategy inspired by the logistic growth differential equation.It employs a gentle decay rate in the initial training stage,a rapid decay rate during the intermediate stage,and a slower decay rate in the network convergence stage.Additionally,the decay rate is adjusted adaptively based on the filter weights at each stage.By controlling the adaptive decay rate at each stage,the pruning of neural network filters can be effectively managed.In experiments conducted on the CIFAR-10 and ILSVRC-2012 datasets,the pruning of neural networks significantly reduces the floating-point operations while maintaining the same pruning rate.Specifically,when implementing a 30%pruning rate on the ResNet-110 network,the pruned neural network not only decreases floating-point operations by 40.8%but also enhances the classification accuracy by 0.49%compared to the original network.展开更多
Memristors have a synapse-like two-terminal structure and electrical properties,which are widely used in the construc-tion of artificial synapses.However,compared to inorganic materials,organic materials are rarely us...Memristors have a synapse-like two-terminal structure and electrical properties,which are widely used in the construc-tion of artificial synapses.However,compared to inorganic materials,organic materials are rarely used for artificial spiking synapses due to their relatively poor memrisitve performance.Here,for the first time,we present an organic memristor based on an electropolymerized dopamine-based memristive layer.This polydopamine-based memristor demonstrates the improve-ments in key performance,including a low threshold voltage of 0.3 V,a thin thickness of 16 nm,and a high parasitic capaci-tance of about 1μF·mm^(-2).By leveraging these properties in combination with its stable threshold switching behavior,we con-struct a capacitor-free and low-power artificial spiking neuron capable of outputting the oscillation voltage,whose spiking fre-quency increases with the increase of current stimulation analogous to a biological neuron.The experimental results indicate that our artificial spiking neuron holds potential for applications in neuromorphic computing and systems.展开更多
With the increasing popularity of blockchain applications, the security of data sources on the blockchain is gradually receiving attention. Providing reliable data for the blockchain safely and efficiently has become ...With the increasing popularity of blockchain applications, the security of data sources on the blockchain is gradually receiving attention. Providing reliable data for the blockchain safely and efficiently has become a research hotspot, and the security of the oracle responsible for providing reliable data has attracted much attention. The most widely used centralized oracles in blockchain, such as Provable and Town Crier, all rely on a single oracle to obtain data, which suffers from a single point of failure and limits the large-scale development of blockchain. To this end, the distributed oracle scheme is put forward, but the existing distributed oracle schemes such as Chainlink and Augur generally have low execution efficiency and high communication overhead, which leads to their poor applicability. To solve the above problems, this paper proposes a trusted distributed oracle scheme based on a share recovery threshold signature. First, a data verification method of distributed oracles is designed based on threshold signature. By aggregating the signatures of oracles, data from different data sources can be mutually verified, leading to a more efficient data verification and aggregation process. Then, a credibility-based cluster head election algorithm is designed, which reduces the communication overhead by clarifying the function distribution and building a hierarchical structure. Considering the good performance of the BLS threshold signature in large-scale applications, this paper combines it with distributed oracle technology and proposes a BLS threshold signature algorithm that supports share recovery in distributed oracles. The share recovery mechanism enables the proposed scheme to solve the key loss issue, and the setting of the threshold value enables the proposed scheme to complete signature aggregation with only a threshold number of oracles, making the scheme more robust. Finally, experimental results indicate that, by using the threshold signature technology and the cluster head election algorithm, our scheme effectively improves the execution efficiency of oracles and solves the problem of a single point of failure, leading to higher scalability and robustness.展开更多
Objective: To study the relationship between cortical auditory evoked potential (CAEP) thresholds and behavioral thresholds in pediatric populations with sensorineural hearing loss (SNHL). Methods: Fifteen children (m...Objective: To study the relationship between cortical auditory evoked potential (CAEP) thresholds and behavioral thresholds in pediatric populations with sensorineural hearing loss (SNHL). Methods: Fifteen children (mean age 6.8 years) with bilateral SNHL underwent behavioral pure-tone audiometry and CAEP testing at 0.5, 1, 2, and 4 kHz. CAEP thresholds were determined using tone bursts, and correlations between CAEP and pure-tone thresholds were analyzed using Pearson correlation and t-tests. Results: A strong positive correlation was observed between P1 thresholds and behavioral thresholds across all test frequencies: 0.5 kHz (r = 0.765, p Conclusion: The strong correlation between P1 and behavioral thresholds demonstrates the reliability of CAEP testing for estimating auditory thresholds in children. These findings support the use of CAEP testing as a reliable objective tool for threshold estimation, particularly in cases where behavioral responses cannot be reliably obtained. When adjusted with frequency-specific correction values, CAEP testing provides a reliable method for assessing hearing thresholds in pediatric populations.展开更多
To better capture the characteristics of asymmetry and structural fluctuations observed in count time series,this study delves into the application of the quantile regression(QR)method for analyzing and forecasting no...To better capture the characteristics of asymmetry and structural fluctuations observed in count time series,this study delves into the application of the quantile regression(QR)method for analyzing and forecasting nonlinear integer-valued time series exhibiting a piecewise phenomenon.Specifically,we focus on the parameter estimation in the first-order Self-Exciting Threshold Integer-valued Autoregressive(SETINAR(2,1))process with symmetry,asymmetry,and contaminated innovations.We establish the asymptotic properties of the estimator under certain regularity conditions.Monte Carlo simulations demonstrate the superior performance of the QR method compared to the conditional least squares(CLS)approach.Furthermore,we validate the robustness of the proposed method through empirical quantile regression estimation and forecasting for larceny incidents and CAD drug call counts in Pittsburgh,showcasing its effectiveness across diverse levels of data heterogeneity.展开更多
The societal risk related to rainfalltriggered rapid debris flows is commonly managed in urbanized areas by means of early warning systems based on monitoring of hydrological parameters(such as rainfall or soil moistu...The societal risk related to rainfalltriggered rapid debris flows is commonly managed in urbanized areas by means of early warning systems based on monitoring of hydrological parameters(such as rainfall or soil moisture) and thresholds values.In Alpine catchments,this type of landslides is recurrent and represent one of the major geohazards.Debris flows are typically initiated by high-intensity rainstorms,prolonged rainfall with moderate intensity or snow melting.They frequently happen in situations of temporary infiltration into soils that are initially unsaturated.During significant rainfall events,the rise in pore water pressure can become crucial for the stability of slopes in particular areas.This phenomenon relies on hydraulic and geotechnical characteristics,along with the thickness of the involved soils.This procedure can result in a local drop in shear strength,as both apparent cohesion and effective stress decline,while driving forces rise because of the increase in unit weight.Accordingly,this study estimates Intensity-Duration(I-D) rainfall thresholds at the site-specific and distributed scales by combining empirical and physics-based approaches and modeling of soil coverings involved in soil slips or debris slides inducing debris flows.The approach was tested for mountain slopes of the Valtellina valley(Lombardia region,northern Italy),which suffered several catastrophic landslide events in the last decades.The empirical approach was adopted to reconstruct physics-based slope models of representative source areas of past debris flows events.To such a scope,nonpunctual but distributed data of hydro-mechanical soil properties and thicknesses were considered.Thus,to reconstruct the unsaturated/saturated critical conditions leading the slope instability,a combined hydrological modeling and infinite-slope stability analysis was adopted.This combined hydromechanical numerical model was used to attempt to determine a three-dimensional Intensity-Duration threshold for landslide initiation considering plausible rainfall for the Valtellina valley.Due to the lack of reliable records of past landslide hindering a thorough empirical analysis,the presented approach can be considered as a feasible approach for establishing a warning standard in urbanized areas at risk of shallow landslides triggered by rainfall.Moreover,findings highlight the importance of having access to spatially distributed soil characteristics to define and enhance input data for physics-based modelling.Finally,the proposed approach can aid an early warning system for the onset of shallow landslides by utilizing real-time rainfall monitoring or now-casting through a meteorological radar technique.展开更多
3D sparse convolution has emerged as a pivotal technique for efficient voxel-based perception in autonomous systems,enabling selective feature extraction from non-empty voxels while suppressing computational waste.Des...3D sparse convolution has emerged as a pivotal technique for efficient voxel-based perception in autonomous systems,enabling selective feature extraction from non-empty voxels while suppressing computational waste.Despite its theoretical efficiency advantages,practical implementations face under-explored limitations:the fixed geometric patterns of conventional sparse convolutional kernels inevitably process non-contributory positions during sliding-window operations,particularly in regions with uneven point cloud density.To address this,we propose Hierarchical Shape Pruning for 3D Sparse Convolution(HSP-S),which dynamically eliminates redundant kernel stripes through layer-adaptive thresholding.Unlike static soft pruning methods,HSP-S maintains trainable sparsity patterns by progressively adjusting pruning thresholds during optimization,enlarging original parameter search space while removing redundant operations.Extensive experiments validate effectiveness of HSP-S acrossmajor autonomous driving benchmarks.On KITTI’s 3D object detection task,our method reduces 93.47%redundant kernel computations whilemaintaining comparable accuracy(1.56%mAP drop).Remarkably,on themore complexNuScenes benchmark,HSP-S achieves simultaneous computation reduction(21.94%sparsity)and accuracy gains(1.02%mAP(mean Average Precision)and 0.47%NDS(nuScenes detection score)improvement),demonstrating its scalability to diverse perception scenarios.This work establishes the first learnable shape pruning framework that simultaneously enhances computational efficiency and preserves detection accuracy in 3D perception systems.展开更多
The dynamic routing mechanism in evolvable networks enables adaptive reconfiguration of topol-ogical structures and transmission pathways based on real-time task requirements and data character-istics.However,the heig...The dynamic routing mechanism in evolvable networks enables adaptive reconfiguration of topol-ogical structures and transmission pathways based on real-time task requirements and data character-istics.However,the heightened architectural complexity and expanded parameter dimensionality in evolvable networks present significant implementation challenges when deployed in resource-con-strained environments.Due to the critical paths ignored,traditional pruning strategies cannot get a desired trade-off between accuracy and efficiency.For this reason,a critical path retention pruning(CPRP)method is proposed.By deeply traversing the computational graph,the dependency rela-tionship among nodes is derived.Then the nodes are grouped and sorted according to their contribu-tion value.The redundant operations are removed as much as possible while ensuring that the criti-cal path is not affected.As a result,computational efficiency is improved while a higher accuracy is maintained.On the CIFAR benchmark,the experimental results demonstrate that CPRP-induced pruning incurs accuracy degradation below 4.00%,while outperforming traditional feature-agnostic grouping methods by an average 8.98%accuracy improvement.Simultaneously,the pruned model attains a 2.41 times inference acceleration while achieving 48.92%parameter compression and 53.40%floating-point operations(FLOPs)reduction.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
Filter pruning effectively compresses the neural network by reducing both its parameters and computational cost.Existing pruning methods typically rely on pre-designed pruning criteria to measure filter importance and...Filter pruning effectively compresses the neural network by reducing both its parameters and computational cost.Existing pruning methods typically rely on pre-designed pruning criteria to measure filter importance and remove those deemed unimportant.However,different layers of the neural network exhibit varying filter distributions,making it inappropriate to implement the same pruning criterion for all layers.Additionally,some approaches apply different criteria from the set of pre-defined pruning rules for different layers,but the limited space leads to the difficulty of covering all layers.If criteria for all layers are manually designed,it is costly and difficult to generalize to other networks.To solve this problem,we present a novel neural network pruning method based on the Criterion Learner and Attention Distillation(CLAD).Specifically,CLAD develops a differentiable criterion learner,which is integrated into each layer of the network.The learner can automatically learn the appropriate pruning criterion according to the filter parameters of each layer,thus the requirement of manual design is eliminated.Furthermore,the criterion learner is trained end-to-end by the gradient optimization algorithm to achieve efficient pruning.In addition,attention distillation,which fully utilizes the knowledge of unpruned networks to guide the optimization of the learner and improve the pruned network performance,is introduced in the process of learner optimization.Experiments conducted on various datasets and networks demonstrate the effectiveness of the proposed method.Notably,CLAD reduces the FLOPs of Res Net-110 by about 53%on the CIFAR-10 dataset,while simultaneously improves the network's accuracy by 0.05%.Moreover,it reduces the FLOPs of Res Net-50 by about 46%on the Image Net-1K dataset,and maintains a top-1 accuracy of 75.45%.展开更多
End-to-end object detection Transformer(DETR)successfully established the paradigm of the Transformer architecture in the field of object detection.Its end-to-end detection process and the idea of set prediction have ...End-to-end object detection Transformer(DETR)successfully established the paradigm of the Transformer architecture in the field of object detection.Its end-to-end detection process and the idea of set prediction have become one of the hottest network architectures in recent years.There has been an abundance of work improving upon DETR.However,DETR and its variants require a substantial amount of memory resources and computational costs,and the vast number of parameters in these networks is unfavorable for model deployment.To address this issue,a greedy pruning(GP)algorithm is proposed,applied to a variant denoising-DETR(DN-DETR),which can eliminate redundant parameters in the Transformer architecture of DN-DETR.Considering the different roles of the multi-head attention(MHA)module and the feed-forward network(FFN)module in the Transformer architecture,a modular greedy pruning(MGP)algorithm is proposed.This algorithm separates the two modules and applies their respective optimal strategies and parameters.The effectiveness of the proposed algorithm is validated on the COCO 2017 dataset.The model obtained through the MGP algorithm reduces the parameters by 49%and the number of floating point operations(FLOPs)by 44%compared to the Transformer architecture of DN-DETR.At the same time,the mean average precision(mAP)of the model increases from 44.1%to 45.3%.展开更多
Radio environment plays an important role in radio astronomy observations.Further analysis is needed on the time and intensity distributions of interference signals for long-term radio environment monitoring.Sample va...Radio environment plays an important role in radio astronomy observations.Further analysis is needed on the time and intensity distributions of interference signals for long-term radio environment monitoring.Sample variance is an important estimate of the interference signal decision threshold.Here,we propose an improved algorithm for calculating data sample variance relying on four established statistical methods:the variance of the trimmed data,winsorized sample variance,median absolute deviation,and median of the trimmed data pairwise averaged squares method.The variance and decision threshold in the protected section of the radio astronomy L-band are calculated.Among the four methods,the improved median of the trimmed data pairwise averaged squares algorithm has higher accuracy,but in a comparison of overall experimental results,the cleanliness rate of all algorithms is above 96%.In a comparison between the improved algorithm and the four methods,the cleanliness rate of the improved algorithm is above 98%,verifying its feasibility.The time-intensity interference distribution in the radio protection band is also obtained.Finally,we use comprehensive monitoring data of radio astronomy protection bands,radio interference bands,and interfered frequency bands to establish a comprehensive evaluation system for radio observatory sites,including the observable time proportion in the radio astronomy protection band,the occasional time-intensity distribution in the radio interference frequency band,and the intensity distribution of the interfered frequency band.展开更多
Low-pressure mercury lamps,with a main emission of about 254 nm,have been used for disinfecting air,water and surfaces for nearly a century.However,only a few studies on the corneal damage threshold at the wavelength ...Low-pressure mercury lamps,with a main emission of about 254 nm,have been used for disinfecting air,water and surfaces for nearly a century.However,only a few studies on the corneal damage threshold at the wavelength of 254nm exist.In this paper,the in vivo corneal damage threshold was determined inchinchilla rabbit model using a laser system at 254 nm.The irradiance of the laser spot was nearly flat-top distributed and the beam diameter on the animal corneal surface was about 3.44mm and 3.28 mm along the horizontal and vertical directions.Damage lesion determinations were performed at 12 h post-exposure using fluorescein sodium staining.TheED50 was17.7 mJ/cm^(2)with a 95%confidence interval of 15.3-20.1 mJ/cm^(2).The obtained results may contribute to the knowledge base for the refinement of the UV hazard function.展开更多
Atmospheric aerosols are the primary contributors to environmental pollution.As such aerosols are micro-to nanosized particles invisible to the naked eye,it is necessary to utilize LiDAR technology for their detection...Atmospheric aerosols are the primary contributors to environmental pollution.As such aerosols are micro-to nanosized particles invisible to the naked eye,it is necessary to utilize LiDAR technology for their detection.The laser radar echo signal is vulnerable to background light and electronic thermal noise.While single-photon LiDAR can effectively reduce background light interference,electronic thermal noise remains a significant challenge,especially at long distances and in environments with a low signal-to-noise ratio(SNR).However,conventional denoising methods cannot achieve satisfactory results in this case.In this paper,a novel adaptive continuous threshold wavelet denoising algorithm is proposed to filter out the noise.The algorithm features an adaptive threshold and a continuous threshold function.The adaptive threshold is dynamically adjusted according to the wavelet decomposition level,and the continuous threshold function ensures continuity with lower constant error,thus optimizing the denoising process.Simulation results show that the proposed algorithm has excellent performance in improving SNR and reducing root mean square error(RMSE)compared with other algorithms.Experimental results show that denoising of an actual LiDAR echo signal results in a 4.37 dB improvement in SNR and a 39.5%reduction in RMSE.The proposed method significantly enhances the ability of single-photon LiDAR to detect weak signals.展开更多
A clock bias data processing method based on interval correlation coefficient wavelet threshold denoising is suggested for minor mistakes in clock bias data in order to increase the efficacy of satellite clock bias pr...A clock bias data processing method based on interval correlation coefficient wavelet threshold denoising is suggested for minor mistakes in clock bias data in order to increase the efficacy of satellite clock bias prediction.Wavelet analysis was first used to break down the satellite clock frequency data into several levels,producing high and low frequency coefficients for each layer.The correlation coefficients of the high and low frequency coefficients in each of the three sub-intervals created by splitting these coefficients were then determined.The major noise region—the sub-interval with the lowest correlation coefficient—was chosen for thresholding treatment and noise threshold computation.The clock frequency data was then processed using wavelet reconstruction and reconverted to clock data.Lastly,three different kinds of satellite clock data—RTS,whu-o,and IGS-F—were used to confirm the produced data.Our method enhanced the stability of the Quadratic Polynomial(QP)model’s predictions for the C16 satellite by about 40%,according to the results.The accuracy and stability of the Auto Regression Integrated Moving Average(ARIMA)model improved up to 41.8%and 14.2%,respectively,whilst the Wavelet Neural Network(WNN)model improved by roughly 27.8%and 63.6%,respectively.Although our method has little effect on forecasting IGS-F series satellites,the experimental findings show that it can improve the accuracy and stability of QP,ARIMA,and WNN model forecasts for RTS and whu-o satellite clock bias.展开更多
In clinical research,subgroup analysis can help identify patient groups that respond better or worse to specific treatments,improve therapeutic effect and safety,and is of great significance in precision medicine.This...In clinical research,subgroup analysis can help identify patient groups that respond better or worse to specific treatments,improve therapeutic effect and safety,and is of great significance in precision medicine.This article considers subgroup analysis methods for longitudinal data containing multiple covariates and biomarkers.We divide subgroups based on whether a linear combination of these biomarkers exceeds a predetermined threshold,and assess the heterogeneity of treatment effects across subgroups using the interaction between subgroups and exposure variables.Quantile regression is used to better characterize the global distribution of the response variable and sparsity penalties are imposed to achieve variable selection of covariates and biomarkers.The effectiveness of our proposed methodology for both variable selection and parameter estimation is verified through random simulations.Finally,we demonstrate the application of this method by analyzing data from the PA.3 trial,further illustrating the practicality of the method proposed in this paper.展开更多
BackgroundIt's crucial to study the effect of changes in thresholds(T)and most comfortable levels(M)on behavioral measurements in young children using cochlear implants.This would help the clinician with the optim...BackgroundIt's crucial to study the effect of changes in thresholds(T)and most comfortable levels(M)on behavioral measurements in young children using cochlear implants.This would help the clinician with the optimization and validation of programming parameters.ObjectiveThe study has attempted to describe the changes in behavioral responses with modification of T and M levels.MethodsTwenty-five participants in the age range 5 to 12 years using HR90K/HiFocus1J or HR90KAdvantage/HiFocus1J with Harmony speech processors participated in the study.A decrease in T levels,a rise in T levels,or a decrease in M levels in the everyday program were used to create experimental programs.Sound field thresholds and speech perception were measured at 50 dBHL for three experimental and everyday programs.ConclusionThe results indicated that only reductions of M levels resulted in significantly(p<0.01)poor aided thresholds and speech perception.On the other hand,variation in T levels did not have significant changes in either sound field thresholds or speech perception.The results highlight that M levels must be correctly established in order to prevent decreased speech perception and audibility.展开更多
基金supported by the National Natural Science Foundation of China,No.31760290,82160688the Key Development Areas Project of Ganzhou Science and Technology,No.2022B-SF9554(all to XL)。
文摘Synaptic pruning is a crucial process in synaptic refinement,eliminating unstable synaptic connections in neural circuits.This process is triggered and regulated primarily by spontaneous neural activity and experience-dependent mechanisms.The pruning process involves multiple molecular signals and a series of regulatory activities governing the“eat me”and“don't eat me”states.Under physiological conditions,the interaction between glial cells and neurons results in the clearance of unnecessary synapses,maintaining normal neural circuit functionality via synaptic pruning.Alterations in genetic and environmental factors can lead to imbalanced synaptic pruning,thus promoting the occurrence and development of autism spectrum disorder,schizophrenia,Alzheimer's disease,and other neurological disorders.In this review,we investigated the molecular mechanisms responsible for synaptic pruning during neural development.We focus on how synaptic pruning can regulate neural circuits and its association with neurological disorders.Furthermore,we discuss the application of emerging optical and imaging technologies to observe synaptic structure and function,as well as their potential for clinical translation.Our aim was to enhance our understanding of synaptic pruning during neural development,including the molecular basis underlying the regulation of synaptic function and the dynamic changes in synaptic density,and to investigate the potential role of these mechanisms in the pathophysiology of neurological diseases,thus providing a theoretical foundation for the treatment of neurological disorders.
基金output of a research project implemented as part of the Basic Research Program at HSE University。
文摘Activation pruning reduces neural network complexity by eliminating low-importance neuron activations,yet identifying the critical pruning threshold—beyond which accuracy rapidly deteriorates—remains computationally expensive and typically requires exhaustive search.We introduce a thermodynamics-inspired framework that treats activation distributions as energy-filtered physical systems and employs the free energy of activations as a principled evaluation metric.Phase-transition-like phenomena in the free-energy profile—such as extrema,inflection points,and curvature changes—yield reliable estimates of the critical pruning threshold,providing a theoretically grounded means of predicting sharp accuracy degradation.To further enhance efficiency,we propose a renormalized free energy technique that approximates full-evaluation free energy using only the activation distribution of the unpruned network.This eliminates repeated forward passes,dramatically reducing computational overhead and achieving speedups of up to 550×for MLPs.Extensive experiments across diverse vision architectures(MLP,CNN,ResNet,MobileNet,Vision Transformer)and text models(LSTM,BERT,ELECTRA,T5,GPT-2)on multiple datasets validate the generality,robustness,and computational efficiency of our approach.Overall,this work establishes a theoretically grounded and practically effective framework for activation pruning,bridging the gap between analytical understanding and efficient deployment of sparse neural networks.
基金supported by a grant(No.CRPG-25-2054)under the Cybersecurity Research and Innovation Pioneers Initiative,provided by the National Cybersecurity Authority(NCA)in the Kingdom of Saudi Arabia.
文摘Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subnetworks in order to mitigate the exposure of sensitive data and reduce the overhead on client devices,thereby making SL particularly suitable for resource-constrained devices.Although SL prevents the direct transmission of raw data,it does not alleviate entirely the risk of privacy breaches.In fact,the data intermediately transmitted to the server sub-model may include patterns or information that could reveal sensitive data.Moreover,achieving a balance between model utility and data privacy has emerged as a challenging problem.In this article,we propose a novel defense approach that combines:(i)Adversarial learning,and(ii)Network channel pruning.In particular,the proposed adversarial learning approach is specifically designed to reduce the risk of private data exposure while maintaining high performance for the utility task.On the other hand,the suggested channel pruning enables the model to adaptively adjust and reactivate pruned channels while conducting adversarial training.The integration of these two techniques reduces the informativeness of the intermediate data transmitted by the client sub-model,thereby enhancing its robustness against attribute inference attacks without adding significant computational overhead,making it wellsuited for IoT devices,mobile platforms,and Internet of Vehicles(IoV)scenarios.The proposed defense approach was evaluated using EfficientNet-B0,a widely adopted compact model,along with three benchmark datasets.The obtained results showcased its superior defense capability against attribute inference attacks compared to existing state-of-the-art methods.This research’s findings demonstrated the effectiveness of the proposed channel pruning-based adversarial training approach in achieving the intended compromise between utility and privacy within SL frameworks.In fact,the classification accuracy attained by the attackers witnessed a drastic decrease of 70%.
基金supported by the National Natural Science Foundation of China under Grant No.62172132.
文摘The surge of large-scale models in recent years has led to breakthroughs in numerous fields,but it has also introduced higher computational costs and more complex network architectures.These increasingly large and intricate networks pose challenges for deployment and execution while also exacerbating the issue of network over-parameterization.To address this issue,various network compression techniques have been developed,such as network pruning.A typical pruning algorithm follows a three-step pipeline involving training,pruning,and retraining.Existing methods often directly set the pruned filters to zero during retraining,significantly reducing the parameter space.However,this direct pruning strategy frequently results in irreversible information loss.In the early stages of training,a network still contains much uncertainty,and evaluating filter importance may not be sufficiently rigorous.To manage the pruning process effectively,this paper proposes a flexible neural network pruning algorithm based on the logistic growth differential equation,considering the characteristics of network training.Unlike other pruning algorithms that directly reduce filter weights,this algorithm introduces a three-stage adaptive weight decay strategy inspired by the logistic growth differential equation.It employs a gentle decay rate in the initial training stage,a rapid decay rate during the intermediate stage,and a slower decay rate in the network convergence stage.Additionally,the decay rate is adjusted adaptively based on the filter weights at each stage.By controlling the adaptive decay rate at each stage,the pruning of neural network filters can be effectively managed.In experiments conducted on the CIFAR-10 and ILSVRC-2012 datasets,the pruning of neural networks significantly reduces the floating-point operations while maintaining the same pruning rate.Specifically,when implementing a 30%pruning rate on the ResNet-110 network,the pruned neural network not only decreases floating-point operations by 40.8%but also enhances the classification accuracy by 0.49%compared to the original network.
基金support from the Beijing Natural Science Foundation-Xiaomi Innovation Joint Fund(No.L233009)National Natural Science Foundation of China(NSFC Nos.62422409,62174152,and 62374159)from the Youth Innovation Promotion Association of Chinese Academy of Sciences(No.2020115).
文摘Memristors have a synapse-like two-terminal structure and electrical properties,which are widely used in the construc-tion of artificial synapses.However,compared to inorganic materials,organic materials are rarely used for artificial spiking synapses due to their relatively poor memrisitve performance.Here,for the first time,we present an organic memristor based on an electropolymerized dopamine-based memristive layer.This polydopamine-based memristor demonstrates the improve-ments in key performance,including a low threshold voltage of 0.3 V,a thin thickness of 16 nm,and a high parasitic capaci-tance of about 1μF·mm^(-2).By leveraging these properties in combination with its stable threshold switching behavior,we con-struct a capacitor-free and low-power artificial spiking neuron capable of outputting the oscillation voltage,whose spiking fre-quency increases with the increase of current stimulation analogous to a biological neuron.The experimental results indicate that our artificial spiking neuron holds potential for applications in neuromorphic computing and systems.
基金supported by the National Natural Science Foundation of China(Grant No.62102449)the Central Plains Talent Program under Grant No.224200510003.
文摘With the increasing popularity of blockchain applications, the security of data sources on the blockchain is gradually receiving attention. Providing reliable data for the blockchain safely and efficiently has become a research hotspot, and the security of the oracle responsible for providing reliable data has attracted much attention. The most widely used centralized oracles in blockchain, such as Provable and Town Crier, all rely on a single oracle to obtain data, which suffers from a single point of failure and limits the large-scale development of blockchain. To this end, the distributed oracle scheme is put forward, but the existing distributed oracle schemes such as Chainlink and Augur generally have low execution efficiency and high communication overhead, which leads to their poor applicability. To solve the above problems, this paper proposes a trusted distributed oracle scheme based on a share recovery threshold signature. First, a data verification method of distributed oracles is designed based on threshold signature. By aggregating the signatures of oracles, data from different data sources can be mutually verified, leading to a more efficient data verification and aggregation process. Then, a credibility-based cluster head election algorithm is designed, which reduces the communication overhead by clarifying the function distribution and building a hierarchical structure. Considering the good performance of the BLS threshold signature in large-scale applications, this paper combines it with distributed oracle technology and proposes a BLS threshold signature algorithm that supports share recovery in distributed oracles. The share recovery mechanism enables the proposed scheme to solve the key loss issue, and the setting of the threshold value enables the proposed scheme to complete signature aggregation with only a threshold number of oracles, making the scheme more robust. Finally, experimental results indicate that, by using the threshold signature technology and the cluster head election algorithm, our scheme effectively improves the execution efficiency of oracles and solves the problem of a single point of failure, leading to higher scalability and robustness.
文摘Objective: To study the relationship between cortical auditory evoked potential (CAEP) thresholds and behavioral thresholds in pediatric populations with sensorineural hearing loss (SNHL). Methods: Fifteen children (mean age 6.8 years) with bilateral SNHL underwent behavioral pure-tone audiometry and CAEP testing at 0.5, 1, 2, and 4 kHz. CAEP thresholds were determined using tone bursts, and correlations between CAEP and pure-tone thresholds were analyzed using Pearson correlation and t-tests. Results: A strong positive correlation was observed between P1 thresholds and behavioral thresholds across all test frequencies: 0.5 kHz (r = 0.765, p Conclusion: The strong correlation between P1 and behavioral thresholds demonstrates the reliability of CAEP testing for estimating auditory thresholds in children. These findings support the use of CAEP testing as a reliable objective tool for threshold estimation, particularly in cases where behavioral responses cannot be reliably obtained. When adjusted with frequency-specific correction values, CAEP testing provides a reliable method for assessing hearing thresholds in pediatric populations.
基金supported by Social Science Planning Foundation of Liaoning Province(Grand No.L22ZD065)National Natural Science Foundation of China(Grand Nos.12271231,1247012719,12001229)。
文摘To better capture the characteristics of asymmetry and structural fluctuations observed in count time series,this study delves into the application of the quantile regression(QR)method for analyzing and forecasting nonlinear integer-valued time series exhibiting a piecewise phenomenon.Specifically,we focus on the parameter estimation in the first-order Self-Exciting Threshold Integer-valued Autoregressive(SETINAR(2,1))process with symmetry,asymmetry,and contaminated innovations.We establish the asymptotic properties of the estimator under certain regularity conditions.Monte Carlo simulations demonstrate the superior performance of the QR method compared to the conditional least squares(CLS)approach.Furthermore,we validate the robustness of the proposed method through empirical quantile regression estimation and forecasting for larceny incidents and CAD drug call counts in Pittsburgh,showcasing its effectiveness across diverse levels of data heterogeneity.
文摘The societal risk related to rainfalltriggered rapid debris flows is commonly managed in urbanized areas by means of early warning systems based on monitoring of hydrological parameters(such as rainfall or soil moisture) and thresholds values.In Alpine catchments,this type of landslides is recurrent and represent one of the major geohazards.Debris flows are typically initiated by high-intensity rainstorms,prolonged rainfall with moderate intensity or snow melting.They frequently happen in situations of temporary infiltration into soils that are initially unsaturated.During significant rainfall events,the rise in pore water pressure can become crucial for the stability of slopes in particular areas.This phenomenon relies on hydraulic and geotechnical characteristics,along with the thickness of the involved soils.This procedure can result in a local drop in shear strength,as both apparent cohesion and effective stress decline,while driving forces rise because of the increase in unit weight.Accordingly,this study estimates Intensity-Duration(I-D) rainfall thresholds at the site-specific and distributed scales by combining empirical and physics-based approaches and modeling of soil coverings involved in soil slips or debris slides inducing debris flows.The approach was tested for mountain slopes of the Valtellina valley(Lombardia region,northern Italy),which suffered several catastrophic landslide events in the last decades.The empirical approach was adopted to reconstruct physics-based slope models of representative source areas of past debris flows events.To such a scope,nonpunctual but distributed data of hydro-mechanical soil properties and thicknesses were considered.Thus,to reconstruct the unsaturated/saturated critical conditions leading the slope instability,a combined hydrological modeling and infinite-slope stability analysis was adopted.This combined hydromechanical numerical model was used to attempt to determine a three-dimensional Intensity-Duration threshold for landslide initiation considering plausible rainfall for the Valtellina valley.Due to the lack of reliable records of past landslide hindering a thorough empirical analysis,the presented approach can be considered as a feasible approach for establishing a warning standard in urbanized areas at risk of shallow landslides triggered by rainfall.Moreover,findings highlight the importance of having access to spatially distributed soil characteristics to define and enhance input data for physics-based modelling.Finally,the proposed approach can aid an early warning system for the onset of shallow landslides by utilizing real-time rainfall monitoring or now-casting through a meteorological radar technique.
文摘3D sparse convolution has emerged as a pivotal technique for efficient voxel-based perception in autonomous systems,enabling selective feature extraction from non-empty voxels while suppressing computational waste.Despite its theoretical efficiency advantages,practical implementations face under-explored limitations:the fixed geometric patterns of conventional sparse convolutional kernels inevitably process non-contributory positions during sliding-window operations,particularly in regions with uneven point cloud density.To address this,we propose Hierarchical Shape Pruning for 3D Sparse Convolution(HSP-S),which dynamically eliminates redundant kernel stripes through layer-adaptive thresholding.Unlike static soft pruning methods,HSP-S maintains trainable sparsity patterns by progressively adjusting pruning thresholds during optimization,enlarging original parameter search space while removing redundant operations.Extensive experiments validate effectiveness of HSP-S acrossmajor autonomous driving benchmarks.On KITTI’s 3D object detection task,our method reduces 93.47%redundant kernel computations whilemaintaining comparable accuracy(1.56%mAP drop).Remarkably,on themore complexNuScenes benchmark,HSP-S achieves simultaneous computation reduction(21.94%sparsity)and accuracy gains(1.02%mAP(mean Average Precision)and 0.47%NDS(nuScenes detection score)improvement),demonstrating its scalability to diverse perception scenarios.This work establishes the first learnable shape pruning framework that simultaneously enhances computational efficiency and preserves detection accuracy in 3D perception systems.
基金Supported by the National Key Research and Development Program of China(No.2022ZD0119003)and the National Natural Science Founda-tion of China(No.61834005).
文摘The dynamic routing mechanism in evolvable networks enables adaptive reconfiguration of topol-ogical structures and transmission pathways based on real-time task requirements and data character-istics.However,the heightened architectural complexity and expanded parameter dimensionality in evolvable networks present significant implementation challenges when deployed in resource-con-strained environments.Due to the critical paths ignored,traditional pruning strategies cannot get a desired trade-off between accuracy and efficiency.For this reason,a critical path retention pruning(CPRP)method is proposed.By deeply traversing the computational graph,the dependency rela-tionship among nodes is derived.Then the nodes are grouped and sorted according to their contribu-tion value.The redundant operations are removed as much as possible while ensuring that the criti-cal path is not affected.As a result,computational efficiency is improved while a higher accuracy is maintained.On the CIFAR benchmark,the experimental results demonstrate that CPRP-induced pruning incurs accuracy degradation below 4.00%,while outperforming traditional feature-agnostic grouping methods by an average 8.98%accuracy improvement.Simultaneously,the pruned model attains a 2.41 times inference acceleration while achieving 48.92%parameter compression and 53.40%floating-point operations(FLOPs)reduction.
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 in part by the National Natural Science Foundation of China under grants 62073085,61973330 and 62350055in part by the Shenzhen Science and Technology Program,China under grant JCYJ20230807093513027in part by the Fundamental Research Funds for the Central Universities,China under grant 1243300008。
文摘Filter pruning effectively compresses the neural network by reducing both its parameters and computational cost.Existing pruning methods typically rely on pre-designed pruning criteria to measure filter importance and remove those deemed unimportant.However,different layers of the neural network exhibit varying filter distributions,making it inappropriate to implement the same pruning criterion for all layers.Additionally,some approaches apply different criteria from the set of pre-defined pruning rules for different layers,but the limited space leads to the difficulty of covering all layers.If criteria for all layers are manually designed,it is costly and difficult to generalize to other networks.To solve this problem,we present a novel neural network pruning method based on the Criterion Learner and Attention Distillation(CLAD).Specifically,CLAD develops a differentiable criterion learner,which is integrated into each layer of the network.The learner can automatically learn the appropriate pruning criterion according to the filter parameters of each layer,thus the requirement of manual design is eliminated.Furthermore,the criterion learner is trained end-to-end by the gradient optimization algorithm to achieve efficient pruning.In addition,attention distillation,which fully utilizes the knowledge of unpruned networks to guide the optimization of the learner and improve the pruned network performance,is introduced in the process of learner optimization.Experiments conducted on various datasets and networks demonstrate the effectiveness of the proposed method.Notably,CLAD reduces the FLOPs of Res Net-110 by about 53%on the CIFAR-10 dataset,while simultaneously improves the network's accuracy by 0.05%.Moreover,it reduces the FLOPs of Res Net-50 by about 46%on the Image Net-1K dataset,and maintains a top-1 accuracy of 75.45%.
基金Shanghai Municipal Commission of Economy and Information Technology,China(No.202301054)。
文摘End-to-end object detection Transformer(DETR)successfully established the paradigm of the Transformer architecture in the field of object detection.Its end-to-end detection process and the idea of set prediction have become one of the hottest network architectures in recent years.There has been an abundance of work improving upon DETR.However,DETR and its variants require a substantial amount of memory resources and computational costs,and the vast number of parameters in these networks is unfavorable for model deployment.To address this issue,a greedy pruning(GP)algorithm is proposed,applied to a variant denoising-DETR(DN-DETR),which can eliminate redundant parameters in the Transformer architecture of DN-DETR.Considering the different roles of the multi-head attention(MHA)module and the feed-forward network(FFN)module in the Transformer architecture,a modular greedy pruning(MGP)algorithm is proposed.This algorithm separates the two modules and applies their respective optimal strategies and parameters.The effectiveness of the proposed algorithm is validated on the COCO 2017 dataset.The model obtained through the MGP algorithm reduces the parameters by 49%and the number of floating point operations(FLOPs)by 44%compared to the Transformer architecture of DN-DETR.At the same time,the mean average precision(mAP)of the model increases from 44.1%to 45.3%.
基金supported by the Ministry of Science and Technology SKA Special Project(2020SKA0110202)the Special Project on Building a Science and Technology Innovation Center for South and Southeast Asia-International Joint Innovation Platform in Yunnan Province:“Yunnan Sino-Malaysian International Joint Laboratory of HF-VHF Advanced Radio Astronomy Technology”(202303AP140003)+4 种基金the National Natural Science Foundation of China(NSFC)Joint Fund for Astronomy(JFA)incubator program(U2031133)the International Partnership Program Project of the International Cooperation Bureau of the Chinese Academy of Sciences:“Belt and Road”Cooperation(114A11KYSB20200001)the Kunming Foreign(International)Cooperation Base Program:“Yunnan Observatory of the Chinese Academy of Sciences-University of Malaya Joint R&D Cooperation Base for Advanced Radio Astronomy Technology”(GHJD-2021022)the China-Malaysia Collaborative Research on Space Remote Sensing and Radio Astronomy Observation of Space Weather at Low and Middle Latitudes under the Key Special Project of the State Key R&D Program of the Ministry of Science and Technol ogy for International Cooperation in Science,Technology and Innovation among Governments(2022YFE0140000)the High-precision calibration method for low-frequency radio interferometric arrays for the SKA project of the Ministry of Science and Technology(2020SKA0110300).
文摘Radio environment plays an important role in radio astronomy observations.Further analysis is needed on the time and intensity distributions of interference signals for long-term radio environment monitoring.Sample variance is an important estimate of the interference signal decision threshold.Here,we propose an improved algorithm for calculating data sample variance relying on four established statistical methods:the variance of the trimmed data,winsorized sample variance,median absolute deviation,and median of the trimmed data pairwise averaged squares method.The variance and decision threshold in the protected section of the radio astronomy L-band are calculated.Among the four methods,the improved median of the trimmed data pairwise averaged squares algorithm has higher accuracy,but in a comparison of overall experimental results,the cleanliness rate of all algorithms is above 96%.In a comparison between the improved algorithm and the four methods,the cleanliness rate of the improved algorithm is above 98%,verifying its feasibility.The time-intensity interference distribution in the radio protection band is also obtained.Finally,we use comprehensive monitoring data of radio astronomy protection bands,radio interference bands,and interfered frequency bands to establish a comprehensive evaluation system for radio observatory sites,including the observable time proportion in the radio astronomy protection band,the occasional time-intensity distribution in the radio interference frequency band,and the intensity distribution of the interfered frequency band.
基金supported by the National Natural Science Foundation of China(No.61575221).
文摘Low-pressure mercury lamps,with a main emission of about 254 nm,have been used for disinfecting air,water and surfaces for nearly a century.However,only a few studies on the corneal damage threshold at the wavelength of 254nm exist.In this paper,the in vivo corneal damage threshold was determined inchinchilla rabbit model using a laser system at 254 nm.The irradiance of the laser spot was nearly flat-top distributed and the beam diameter on the animal corneal surface was about 3.44mm and 3.28 mm along the horizontal and vertical directions.Damage lesion determinations were performed at 12 h post-exposure using fluorescein sodium staining.TheED50 was17.7 mJ/cm^(2)with a 95%confidence interval of 15.3-20.1 mJ/cm^(2).The obtained results may contribute to the knowledge base for the refinement of the UV hazard function.
基金funded by the National Key R&D Program of China(Grant No.2022YFC3300705)the National Natural Science Foundation of China(Grant Nos.62203056,12202048,and 62201056).
文摘Atmospheric aerosols are the primary contributors to environmental pollution.As such aerosols are micro-to nanosized particles invisible to the naked eye,it is necessary to utilize LiDAR technology for their detection.The laser radar echo signal is vulnerable to background light and electronic thermal noise.While single-photon LiDAR can effectively reduce background light interference,electronic thermal noise remains a significant challenge,especially at long distances and in environments with a low signal-to-noise ratio(SNR).However,conventional denoising methods cannot achieve satisfactory results in this case.In this paper,a novel adaptive continuous threshold wavelet denoising algorithm is proposed to filter out the noise.The algorithm features an adaptive threshold and a continuous threshold function.The adaptive threshold is dynamically adjusted according to the wavelet decomposition level,and the continuous threshold function ensures continuity with lower constant error,thus optimizing the denoising process.Simulation results show that the proposed algorithm has excellent performance in improving SNR and reducing root mean square error(RMSE)compared with other algorithms.Experimental results show that denoising of an actual LiDAR echo signal results in a 4.37 dB improvement in SNR and a 39.5%reduction in RMSE.The proposed method significantly enhances the ability of single-photon LiDAR to detect weak signals.
基金2023 Liaoning Institute of Science and Technology Doctoral Program Launch fund(No.2307B29).
文摘A clock bias data processing method based on interval correlation coefficient wavelet threshold denoising is suggested for minor mistakes in clock bias data in order to increase the efficacy of satellite clock bias prediction.Wavelet analysis was first used to break down the satellite clock frequency data into several levels,producing high and low frequency coefficients for each layer.The correlation coefficients of the high and low frequency coefficients in each of the three sub-intervals created by splitting these coefficients were then determined.The major noise region—the sub-interval with the lowest correlation coefficient—was chosen for thresholding treatment and noise threshold computation.The clock frequency data was then processed using wavelet reconstruction and reconverted to clock data.Lastly,three different kinds of satellite clock data—RTS,whu-o,and IGS-F—were used to confirm the produced data.Our method enhanced the stability of the Quadratic Polynomial(QP)model’s predictions for the C16 satellite by about 40%,according to the results.The accuracy and stability of the Auto Regression Integrated Moving Average(ARIMA)model improved up to 41.8%and 14.2%,respectively,whilst the Wavelet Neural Network(WNN)model improved by roughly 27.8%and 63.6%,respectively.Although our method has little effect on forecasting IGS-F series satellites,the experimental findings show that it can improve the accuracy and stability of QP,ARIMA,and WNN model forecasts for RTS and whu-o satellite clock bias.
基金Supported by the Natural Science Foundation of Fujian Province(2022J011177,2024J01903)the Key Project of Fujian Provincial Education Department(JZ230054)。
文摘In clinical research,subgroup analysis can help identify patient groups that respond better or worse to specific treatments,improve therapeutic effect and safety,and is of great significance in precision medicine.This article considers subgroup analysis methods for longitudinal data containing multiple covariates and biomarkers.We divide subgroups based on whether a linear combination of these biomarkers exceeds a predetermined threshold,and assess the heterogeneity of treatment effects across subgroups using the interaction between subgroups and exposure variables.Quantile regression is used to better characterize the global distribution of the response variable and sparsity penalties are imposed to achieve variable selection of covariates and biomarkers.The effectiveness of our proposed methodology for both variable selection and parameter estimation is verified through random simulations.Finally,we demonstrate the application of this method by analyzing data from the PA.3 trial,further illustrating the practicality of the method proposed in this paper.
文摘BackgroundIt's crucial to study the effect of changes in thresholds(T)and most comfortable levels(M)on behavioral measurements in young children using cochlear implants.This would help the clinician with the optimization and validation of programming parameters.ObjectiveThe study has attempted to describe the changes in behavioral responses with modification of T and M levels.MethodsTwenty-five participants in the age range 5 to 12 years using HR90K/HiFocus1J or HR90KAdvantage/HiFocus1J with Harmony speech processors participated in the study.A decrease in T levels,a rise in T levels,or a decrease in M levels in the everyday program were used to create experimental programs.Sound field thresholds and speech perception were measured at 50 dBHL for three experimental and everyday programs.ConclusionThe results indicated that only reductions of M levels resulted in significantly(p<0.01)poor aided thresholds and speech perception.On the other hand,variation in T levels did not have significant changes in either sound field thresholds or speech perception.The results highlight that M levels must be correctly established in order to prevent decreased speech perception and audibility.
