Stochastic neural networks are usually built by introducing random fluctuations into the network. A natural method is to use stochastic connections rather than stochastic activation functions. We propose a new model i...Stochastic neural networks are usually built by introducing random fluctuations into the network. A natural method is to use stochastic connections rather than stochastic activation functions. We propose a new model in which each neuron has very simple functionality but all the connections are stochastic. It is shown that the stationary distribution of the network uniquely exists and it is approxi-mately a Boltzmann-Gibbs distribution. The relationship between the model and the Markov random field is discussed. New techniques to implement simulated annealing and Boltzmann learning are pro-posed. Simulation results on the graph bisection problem and image recognition show that the network is powerful enough to solve real world problems.展开更多
Binary sensor network(BSN) are becoming more attractive due to the low cost deployment,small size,low energy consumption and simple operation.There are two different ways for target tracking in BSN,the weighted algori...Binary sensor network(BSN) are becoming more attractive due to the low cost deployment,small size,low energy consumption and simple operation.There are two different ways for target tracking in BSN,the weighted algorithms and particle filtering algorithm.The weighted algorithms have good realtime property,however have poor estimation property and some of them does not suit for target’s variable velocity model.The particle filtering algorithm can estimate target's position more accurately with poor realtime property and is not suitable for target’s constant velocity model.In this paper distance weight is adopted to estimate the target’s position,which is different from the existing distance weight in other papers.On the analysis of principle of distance weight (DW),prediction-based distance weighted(PDW) algorithm for target tracking in BSN is proposed.Simulation results proved PDW fits for target's constant and variable velocity models with accurate estimation and good realtime property.展开更多
A novel pitch determination algorithm with binary lateral inhibition network (BLIN) is proposed in this paper. A thread like spectrum, which is composed of the fundamental frequency and its harmonic components, is a...A novel pitch determination algorithm with binary lateral inhibition network (BLIN) is proposed in this paper. A thread like spectrum, which is composed of the fundamental frequency and its harmonic components, is acquired by applying a BLIN to the short time spectrum of the speech signal. Then the pitch is determined by the average interval of harmonics. The algorithm is evaluated on COSDIC speech database. For comparison, the same results obtained from the same speech sample with the cepstrum and autocorrelation based pitch determination algorithms are also presented. The results show that the new algorithm is superior to the cepstrum and autocorrelation based pitch determination algorithms.展开更多
The aircraft condition monitoring network is responsible for collecting the status of each component in aircraft. The reliability of this network has a significant effect on safety of the aircraft. The aircraft condit...The aircraft condition monitoring network is responsible for collecting the status of each component in aircraft. The reliability of this network has a significant effect on safety of the aircraft. The aircraft condition monitoring network works in a real-time manner that all the data should be transmitted within the deadline to ensure that the control center makes proper decision in time. Only the connectedness between the source node and destination cannot guarantee the data to be transmitted in time. In this paper, we take the time deadline into account and build the task-based reliability model. The binary decision diagram （BDD）, which has the merit of efficiency in computing and storage space, is introduced when calculating the reliability of the network and addressing the essential variable. A case is analyzed using the algorithm proposed in this paper. The experimental results show that our method is efficient and proper for the reliability analysis of the real-time network.展开更多
Traditional classification methods for cervical cells heavily rely on manual feature extraction,constraining their versatility due to the intricacies of cytology images.Although deep learning approaches offer remarkab...Traditional classification methods for cervical cells heavily rely on manual feature extraction,constraining their versatility due to the intricacies of cytology images.Although deep learning approaches offer remarkable po-tential,they often sacrifice domain-specific knowledge,particularly the morphological patterns characterizing various cell subtypes during automated feature extraction.To bridge this gap,we introduce a novel hierarchical framework that integrates robust features from color,texture,and morphology with latent representations discovered by an improved attention-based multi-scale local binary convolutional neural networks(MS-LBCNN),designed to facilitate powerful feature extraction mechanism.We enhance the standard 6-class Bethesda system(TBS)classification by incorporating a coarse-to-refine fusion strategy,which optimizes the classification pro-cess.The proposed method is uniquely equipped to manage the complexities present in both individual and clustered cell images.Upon rigorous evaluation across three independent data cohorts,our method consistently surpassed existing state-of-the-art techniques.The experimental results indicated the potential of our method in enhancing the development of automation-aided diagnostic systems,and bolstering both the accuracy and ef-ficiency of cytology screening procedures.展开更多
Binary neural networks(BNNs)show promising utilization in cost and power-restricted domains such as edge devices and mobile systems.This is due to its significantly less computation and storage demand,but at the cost ...Binary neural networks(BNNs)show promising utilization in cost and power-restricted domains such as edge devices and mobile systems.This is due to its significantly less computation and storage demand,but at the cost of degraded performance.To close the accuracy gap,in this paper we propose to add a complementary activation function(AF)ahead of the sign based binarization,and rely on the genetic algorithm(GA)to automatically search for the ideal AFs.These AFs can help extract extra information from the input data in the forward pass,while allowing improved gradient approximation in the backward pass.Fifteen novel AFs are identified through our GA-based search,while most of them show improved performance(up to 2.54%on ImageNet)when testing on different datasets and network models.Interestingly,periodic functions are identified as a key component for most of the discovered AFs,which rarely exist in human designed AFs.Our method offers a novel approach for designing general and application-specific BNN architecture.GAAF will be released on GitHub.展开更多
For name-based routing/switching in NDN, the key challenges are to manage large-scale forwarding Tables, to lookup long names of variable lengths, and to deal with frequent updates. Hashing associated with proper leng...For name-based routing/switching in NDN, the key challenges are to manage large-scale forwarding Tables, to lookup long names of variable lengths, and to deal with frequent updates. Hashing associated with proper length-detecting is a straightforward yet efficient solution. Binary search strategy can reduce the number of required hash detecting in the worst case. However, to assure the searching path correct in such a schema, either backtrack searching or redundantly storing some prefixes is required, leading to performance or memory issues as a result. In this paper, we make a deep study on the binary search, and propose a novel mechanism to ensure correct searching path without neither additional backtrack costs nor redundant memory consumptions. Along any binary search path, a bloom filter is employed at each branching point to verify whether a said prefix is present, instead of storing that prefix here. By this means, we can gain significantly optimization on memory efficiency, at the cost of bloom checking before each detecting. Our evaluation experiments on both real-world and randomly synthesized data sets demonstrate our superiorities clearly展开更多
Resistive random access memory(RRAM)has been demonstrated to implement multiply-and-accumulate(MAC)operations using a highly parallel analog fashion,which dramatically accelerates the convolutional neural networks(CNN...Resistive random access memory(RRAM)has been demonstrated to implement multiply-and-accumulate(MAC)operations using a highly parallel analog fashion,which dramatically accelerates the convolutional neural networks(CNNs).Since CNNs require considerable converters between analog crossbars and digital peripheral circuits,recent studies map the binary neural networks(BNNs)onto RRAM and binarize the weights to{+1,-1}.However,two mainstream representations for BNN weights introduce patterns of redundant 0s and 1s when dealing with negative weights.In this work,we reduce the area of redundant 0s and 1s by proposing a BNN weight representation framework based on the novel pattern representation and a corresponding architecture.First,we spilt the weight matrix into several small matrices by clustering adjacent columns together.Second,we extract 1s'patterns,i.e.,the submatrices only containing 1s,from the small weight matrix,such that each final output can be represented by the sum of several patterns.Third,we map these patterns onto RRAM crossbars,including pattern computation crossbars(PCCs)and pattern accumulation crossbars(PACs).Finally,we compare the pattern representation with two mainstream representations and adopt the more area efficient one.The evaluation results demonstrate that our framework can save over 20%of crossbar area effectively,compared with two mainstream representations.展开更多
This paper deals with distributed state estimation problem for discrete time-varying systems over binary sensor networks,where every binary sensor is equipped with an energy harvester.The input of every binary sensor ...This paper deals with distributed state estimation problem for discrete time-varying systems over binary sensor networks,where every binary sensor is equipped with an energy harvester.The input of every binary sensor considers the randomly occurring missing measurements.The differences between the real and estimated inputs of binary sensor are employed to derive useful information in order to address the insufficient information for estimation purpose.The information from neighboring nodes is transmitted only if its energy level is positive,where a random variable is introduced to formulate the energy level.By means of the available information,distributed estimator is constructed for each binary sensor and the desirable performance constraints is given for the dynamic characteristics of estimation errors within anite time horizon.Sucient conditions are established for the existence of desired distribution estimation quantities through local performance analysis methods.Also,the desired distributed estimator gains are calculated recursively,which means the desirable scalability.Ultimately,the viability and efficiency of the distributed scheme are exhibited through a practical illustration.展开更多
Stretchable electronics are in high demand for next-generation wearable devices, but their fabrication is still challenging. Stretchable conductors, flexible pressure sensors, and foldable light-emitting diodes （LEDs...Stretchable electronics are in high demand for next-generation wearable devices, but their fabrication is still challenging. Stretchable conductors, flexible pressure sensors, and foldable light-emitting diodes （LEDs） have been reported; however, the fabrication of stable stretchable batteries, as power suppliers for wearable devices, is significantly behind the development of other stretchable electronics. Several stretchable lithium-ion batteries and primary batteries have been fabricated, but their low capacities and complicated manufacturing processes are obstacles for practical applications. Herein, we report a stretchable zinc/manganese-oxide （Zn-MnO2） full battery based on a silver-nanowire- coated sponge prepared via a facile dip-coating process. The spongy electrode, with a three-dimensional （3D） binary network structure, provided not only high conductivity and stretchability, but also enabled a high mass loading of electrochemically active materials （Zn and MnO2 particles）. The fabricated Zn-MnO2 battery exhibited an areal capacity as high as 3.6 mAh·cm^-2 and could accommodate tensile strains of up to 100% while retaining 89% of its original capacity. The facile solution-based strategy of dip-coating active materials onto a cheap sponge-based stretchable current collector opens up a new avenue for fabricating stretchable batteries.展开更多
Binary mixtures of irregular materials of different particle sizes and/or particle densities are fluidized in a 15-cm diameter column with a perforated plate distributor. An attempt has been made in this work to deter...Binary mixtures of irregular materials of different particle sizes and/or particle densities are fluidized in a 15-cm diameter column with a perforated plate distributor. An attempt has been made in this work to determine the segregation characteristics of jetsam particles for both the homogeneous and heterogeneous binary mixtures in terms of segregation distance by correlating it to the various system parameters, viz. initial static bed height, height of a layer of particles above the bottom grid, superficial gas velocity and average particle size and/or particle densities of the mixture through the dimensional analysis. Correlation on the basis of Artificial Neural Network approach has also been developed with the above system parameters thereby authenticating the development of correlation by the former approach. The calculated values of the segregation distance obtained for both the homogeneous and heterogeneous binary mixtures from both the types of ftuidized beds （i.e. under the static bed condition and the ftuidized bed condition） have also been compared with each other.展开更多
基金This work was supported by the National Natural Science Foundation of China (Grant No. NSFC69805002) the Natural Science Foundation of Zhejiang Province, the Ministry of Eductation Grant for Excellent Youth the Ningbo Science Foundation for Youth (G
文摘Stochastic neural networks are usually built by introducing random fluctuations into the network. A natural method is to use stochastic connections rather than stochastic activation functions. We propose a new model in which each neuron has very simple functionality but all the connections are stochastic. It is shown that the stationary distribution of the network uniquely exists and it is approxi-mately a Boltzmann-Gibbs distribution. The relationship between the model and the Markov random field is discussed. New techniques to implement simulated annealing and Boltzmann learning are pro-posed. Simulation results on the graph bisection problem and image recognition show that the network is powerful enough to solve real world problems.
基金This work is supported by The National Science Fund for Distinguished Young Scholars （60725105） National Basic Research Program of China （973 Program） （2009CB320404）+5 种基金 Program for Changjiang Scholars and Innovative Research Team in University （IRT0852） The National Natural Science Foundation of China （60972048, 61072068） The Special Fund of State Key Laboratory （ISN01080301） The Major program of National Science and Technology （2009ZX03007- 004） Supported by the 111 Project （B08038） The Key Project of Chinese Ministry of Education （107103）.
文摘Binary sensor network(BSN) are becoming more attractive due to the low cost deployment,small size,low energy consumption and simple operation.There are two different ways for target tracking in BSN,the weighted algorithms and particle filtering algorithm.The weighted algorithms have good realtime property,however have poor estimation property and some of them does not suit for target’s variable velocity model.The particle filtering algorithm can estimate target's position more accurately with poor realtime property and is not suitable for target’s constant velocity model.In this paper distance weight is adopted to estimate the target’s position,which is different from the existing distance weight in other papers.On the analysis of principle of distance weight (DW),prediction-based distance weighted(PDW) algorithm for target tracking in BSN is proposed.Simulation results proved PDW fits for target's constant and variable velocity models with accurate estimation and good realtime property.
文摘A novel pitch determination algorithm with binary lateral inhibition network (BLIN) is proposed in this paper. A thread like spectrum, which is composed of the fundamental frequency and its harmonic components, is acquired by applying a BLIN to the short time spectrum of the speech signal. Then the pitch is determined by the average interval of harmonics. The algorithm is evaluated on COSDIC speech database. For comparison, the same results obtained from the same speech sample with the cepstrum and autocorrelation based pitch determination algorithms are also presented. The results show that the new algorithm is superior to the cepstrum and autocorrelation based pitch determination algorithms.
基金National Natural Science Foundation of China (60879024)
文摘The aircraft condition monitoring network is responsible for collecting the status of each component in aircraft. The reliability of this network has a significant effect on safety of the aircraft. The aircraft condition monitoring network works in a real-time manner that all the data should be transmitted within the deadline to ensure that the control center makes proper decision in time. Only the connectedness between the source node and destination cannot guarantee the data to be transmitted in time. In this paper, we take the time deadline into account and build the task-based reliability model. The binary decision diagram （BDD）, which has the merit of efficiency in computing and storage space, is introduced when calculating the reliability of the network and addressing the essential variable. A case is analyzed using the algorithm proposed in this paper. The experimental results show that our method is efficient and proper for the reliability analysis of the real-time network.
基金supported by the Beijing Capital Health Development Research Project[Grant no.2024-2-1031]the Beijing Municipal Natural Science Foundation[Grant no.7192105].
文摘Traditional classification methods for cervical cells heavily rely on manual feature extraction,constraining their versatility due to the intricacies of cytology images.Although deep learning approaches offer remarkable po-tential,they often sacrifice domain-specific knowledge,particularly the morphological patterns characterizing various cell subtypes during automated feature extraction.To bridge this gap,we introduce a novel hierarchical framework that integrates robust features from color,texture,and morphology with latent representations discovered by an improved attention-based multi-scale local binary convolutional neural networks(MS-LBCNN),designed to facilitate powerful feature extraction mechanism.We enhance the standard 6-class Bethesda system(TBS)classification by incorporating a coarse-to-refine fusion strategy,which optimizes the classification pro-cess.The proposed method is uniquely equipped to manage the complexities present in both individual and clustered cell images.Upon rigorous evaluation across three independent data cohorts,our method consistently surpassed existing state-of-the-art techniques.The experimental results indicated the potential of our method in enhancing the development of automation-aided diagnostic systems,and bolstering both the accuracy and ef-ficiency of cytology screening procedures.
文摘Binary neural networks(BNNs)show promising utilization in cost and power-restricted domains such as edge devices and mobile systems.This is due to its significantly less computation and storage demand,but at the cost of degraded performance.To close the accuracy gap,in this paper we propose to add a complementary activation function(AF)ahead of the sign based binarization,and rely on the genetic algorithm(GA)to automatically search for the ideal AFs.These AFs can help extract extra information from the input data in the forward pass,while allowing improved gradient approximation in the backward pass.Fifteen novel AFs are identified through our GA-based search,while most of them show improved performance(up to 2.54%on ImageNet)when testing on different datasets and network models.Interestingly,periodic functions are identified as a key component for most of the discovered AFs,which rarely exist in human designed AFs.Our method offers a novel approach for designing general and application-specific BNN architecture.GAAF will be released on GitHub.
基金supported by the National Natural Science Foundation of China (Grant No. 61472130 and 61702174)the China Postdoctoral Science Foundation funded project
文摘For name-based routing/switching in NDN, the key challenges are to manage large-scale forwarding Tables, to lookup long names of variable lengths, and to deal with frequent updates. Hashing associated with proper length-detecting is a straightforward yet efficient solution. Binary search strategy can reduce the number of required hash detecting in the worst case. However, to assure the searching path correct in such a schema, either backtrack searching or redundantly storing some prefixes is required, leading to performance or memory issues as a result. In this paper, we make a deep study on the binary search, and propose a novel mechanism to ensure correct searching path without neither additional backtrack costs nor redundant memory consumptions. Along any binary search path, a bloom filter is employed at each branching point to verify whether a said prefix is present, instead of storing that prefix here. By this means, we can gain significantly optimization on memory efficiency, at the cost of bloom checking before each detecting. Our evaluation experiments on both real-world and randomly synthesized data sets demonstrate our superiorities clearly
基金partly supported by the National Key Research and Development Program of China under Grant No.2020AAA0130400Beijing Municipal Science and Technology Program of China under Grant No.Z201100004220007+2 种基金the National Natural Science Foundation of China under Grant No.62090021Beijing Academy of Artificial Intelligence(BAAI)Alibaba Innovative Research(AIR)Program.
文摘Resistive random access memory(RRAM)has been demonstrated to implement multiply-and-accumulate(MAC)operations using a highly parallel analog fashion,which dramatically accelerates the convolutional neural networks(CNNs).Since CNNs require considerable converters between analog crossbars and digital peripheral circuits,recent studies map the binary neural networks(BNNs)onto RRAM and binarize the weights to{+1,-1}.However,two mainstream representations for BNN weights introduce patterns of redundant 0s and 1s when dealing with negative weights.In this work,we reduce the area of redundant 0s and 1s by proposing a BNN weight representation framework based on the novel pattern representation and a corresponding architecture.First,we spilt the weight matrix into several small matrices by clustering adjacent columns together.Second,we extract 1s'patterns,i.e.,the submatrices only containing 1s,from the small weight matrix,such that each final output can be represented by the sum of several patterns.Third,we map these patterns onto RRAM crossbars,including pattern computation crossbars(PCCs)and pattern accumulation crossbars(PACs).Finally,we compare the pattern representation with two mainstream representations and adopt the more area efficient one.The evaluation results demonstrate that our framework can save over 20%of crossbar area effectively,compared with two mainstream representations.
基金supported in part by the National Natural Science Foundation of China under Grants 62073070 and U21A2019,and in part by the Alexander von Humboldt Foundation of Germany.
文摘This paper deals with distributed state estimation problem for discrete time-varying systems over binary sensor networks,where every binary sensor is equipped with an energy harvester.The input of every binary sensor considers the randomly occurring missing measurements.The differences between the real and estimated inputs of binary sensor are employed to derive useful information in order to address the insufficient information for estimation purpose.The information from neighboring nodes is transmitted only if its energy level is positive,where a random variable is introduced to formulate the energy level.By means of the available information,distributed estimator is constructed for each binary sensor and the desirable performance constraints is given for the dynamic characteristics of estimation errors within anite time horizon.Sucient conditions are established for the existence of desired distribution estimation quantities through local performance analysis methods.Also,the desired distributed estimator gains are calculated recursively,which means the desirable scalability.Ultimately,the viability and efficiency of the distributed scheme are exhibited through a practical illustration.
基金We acknowledge the funding support from the National Natural Science Foundation of China （Nos. 21431006 and 21761132008）, the Foundation for Innovative Research Groups of the National Natural Science Foundation of China （No. 21521001）, Key Research Program of Frontier Sciences, CAS （No. QYZDJ-SSW- SLH036）, the National Basic Research Program of China （No. 2014CB931800）, and the Users with Excellence and Scientific Research Grant of Hefei Science Center of CAS （No. 2015HSC-UE007）.
文摘Stretchable electronics are in high demand for next-generation wearable devices, but their fabrication is still challenging. Stretchable conductors, flexible pressure sensors, and foldable light-emitting diodes （LEDs） have been reported; however, the fabrication of stable stretchable batteries, as power suppliers for wearable devices, is significantly behind the development of other stretchable electronics. Several stretchable lithium-ion batteries and primary batteries have been fabricated, but their low capacities and complicated manufacturing processes are obstacles for practical applications. Herein, we report a stretchable zinc/manganese-oxide （Zn-MnO2） full battery based on a silver-nanowire- coated sponge prepared via a facile dip-coating process. The spongy electrode, with a three-dimensional （3D） binary network structure, provided not only high conductivity and stretchability, but also enabled a high mass loading of electrochemically active materials （Zn and MnO2 particles）. The fabricated Zn-MnO2 battery exhibited an areal capacity as high as 3.6 mAh·cm^-2 and could accommodate tensile strains of up to 100% while retaining 89% of its original capacity. The facile solution-based strategy of dip-coating active materials onto a cheap sponge-based stretchable current collector opens up a new avenue for fabricating stretchable batteries.
文摘Binary mixtures of irregular materials of different particle sizes and/or particle densities are fluidized in a 15-cm diameter column with a perforated plate distributor. An attempt has been made in this work to determine the segregation characteristics of jetsam particles for both the homogeneous and heterogeneous binary mixtures in terms of segregation distance by correlating it to the various system parameters, viz. initial static bed height, height of a layer of particles above the bottom grid, superficial gas velocity and average particle size and/or particle densities of the mixture through the dimensional analysis. Correlation on the basis of Artificial Neural Network approach has also been developed with the above system parameters thereby authenticating the development of correlation by the former approach. The calculated values of the segregation distance obtained for both the homogeneous and heterogeneous binary mixtures from both the types of ftuidized beds （i.e. under the static bed condition and the ftuidized bed condition） have also been compared with each other.
