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
针对现有高效用项集挖掘算法存在的阈值提升缓慢、剪枝效用差等问题,提出了一种能够更加高效地挖掘效用值最大的前k个项集的算法。TKUL(minging Top-K high Utility itemsets based List)算法综合采用RIUQ、CUDQ和EPB阈值提升策略,加快...针对现有高效用项集挖掘算法存在的阈值提升缓慢、剪枝效用差等问题,提出了一种能够更加高效地挖掘效用值最大的前k个项集的算法。TKUL(minging Top-K high Utility itemsets based List)算法综合采用RIUQ、CUDQ和EPB阈值提升策略,加快最小阈值获取的速度,大大减少了生成的非高效用项集的数量,并通过RUI和EUCPM策略进行剪枝,有效缩小了搜索空间的规模,从而提高了高效用项集的挖掘效率。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
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
Accurate traffic flow prediction(TFP)is vital for efficient and sustainable transportation management and the development of intelligent traffic systems.However,missing data in real-world traffic datasets poses a sign...Accurate traffic flow prediction(TFP)is vital for efficient and sustainable transportation management and the development of intelligent traffic systems.However,missing data in real-world traffic datasets poses a significant challenge to maintaining prediction precision.This study introduces REPTF-TMDI,a novel method that combines a Reduced Error Pruning Tree Forest(REPTree Forest)with a newly proposed Time-based Missing Data Imputation(TMDI)approach.The REP Tree Forest,an ensemble learning approach,is tailored for time-related traffic data to enhance predictive accuracy and support the evolution of sustainable urbanmobility solutions.Meanwhile,the TMDI approach exploits temporal patterns to estimate missing values reliably whenever empty fields are encountered.The proposed method was evaluated using hourly traffic flow data from a major U.S.roadway spanning 2012-2018,incorporating temporal features(e.g.,hour,day,month,year,weekday),holiday indicator,and weather conditions(temperature,rain,snow,and cloud coverage).Experimental results demonstrated that the REPTF-TMDI method outperformed conventional imputation techniques across various missing data ratios by achieving an average 11.76%improvement in terms of correlation coefficient(R).Furthermore,REPTree Forest achieved improvements of 68.62%in RMSE and 70.52%in MAE compared to existing state-of-the-art models.These findings highlight the method’s ability to significantly boost traffic flow prediction accuracy,even in the presence of missing data,thereby contributing to the broader objectives of sustainable urban transportation systems.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.
文摘针对现有高效用项集挖掘算法存在的阈值提升缓慢、剪枝效用差等问题,提出了一种能够更加高效地挖掘效用值最大的前k个项集的算法。TKUL(minging Top-K high Utility itemsets based List)算法综合采用RIUQ、CUDQ和EPB阈值提升策略,加快最小阈值获取的速度,大大减少了生成的非高效用项集的数量,并通过RUI和EUCPM策略进行剪枝,有效缩小了搜索空间的规模,从而提高了高效用项集的挖掘效率。
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%.
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.
文摘Accurate traffic flow prediction(TFP)is vital for efficient and sustainable transportation management and the development of intelligent traffic systems.However,missing data in real-world traffic datasets poses a significant challenge to maintaining prediction precision.This study introduces REPTF-TMDI,a novel method that combines a Reduced Error Pruning Tree Forest(REPTree Forest)with a newly proposed Time-based Missing Data Imputation(TMDI)approach.The REP Tree Forest,an ensemble learning approach,is tailored for time-related traffic data to enhance predictive accuracy and support the evolution of sustainable urbanmobility solutions.Meanwhile,the TMDI approach exploits temporal patterns to estimate missing values reliably whenever empty fields are encountered.The proposed method was evaluated using hourly traffic flow data from a major U.S.roadway spanning 2012-2018,incorporating temporal features(e.g.,hour,day,month,year,weekday),holiday indicator,and weather conditions(temperature,rain,snow,and cloud coverage).Experimental results demonstrated that the REPTF-TMDI method outperformed conventional imputation techniques across various missing data ratios by achieving an average 11.76%improvement in terms of correlation coefficient(R).Furthermore,REPTree Forest achieved improvements of 68.62%in RMSE and 70.52%in MAE compared to existing state-of-the-art models.These findings highlight the method’s ability to significantly boost traffic flow prediction accuracy,even in the presence of missing data,thereby contributing to the broader objectives of sustainable urban transportation systems.
