The proliferation of deep learning(DL)has amplified the demand for processing large and complex datasets for tasks such as modeling,classification,and ***,traditional DL methods compromise client privacy by collecting...
详细信息
The proliferation of deep learning(DL)has amplified the demand for processing large and complex datasets for tasks such as modeling,classification,and ***,traditional DL methods compromise client privacy by collecting sensitive data,underscoring the necessity for privacy-preserving solutions like Federated Learning(FL).FL effectively addresses escalating privacy concerns by facilitating collaborative model training without necessitating the sharing of raw *** that FL clients autonomously manage training data,encouraging client engagement is pivotal for successful model *** overcome challenges like unreliable communication and budget constraints,we present ENTIRE,a contract-based dynamic participation incentive mechanism for *** ensures impartial model training by tailoring participation levels and payments to accommodate diverse client *** approach involves several key ***,we examine how random client participation impacts FL convergence in non-convex scenarios,establishing the correlation between client participation levels and model ***,we reframe model performance optimization as an optimal contract design challenge to guide the distribution of rewards among clients with varying participation *** balancing budget considerations with model effectiveness,we craft optimal contracts for different budgetary constraints,prompting clients to disclose their participation preferences and select suitable contracts for contributing to model ***,we conduct a comprehensive experimental evaluation of ENTIRE using three real *** results demonstrate a significant 12.9%enhancement in model performance,validating its adherence to anticipated economic properties.
Self-supervised learning (SSL) automatically generates internal labels by exploring the potential auxiliary task of the network itself, and trains the same model through these annotations to learn the latent represent...
详细信息
Self-supervised learning (SSL) automatically generates internal labels by exploring the potential auxiliary task of the network itself, and trains the same model through these annotations to learn the latent representation of the data, which greatly improves the accuracy of object detection in remote sensing images (RSIs). However, most existing methods suffer from guaranteeing the quality of the generated SSL pseudo annotations, and the constructed auxiliary tasks are not detectionoriented, which is difficult to enhance the feature representations that are beneficial for object detection. In this paper, we focus on generating more accurate constraints by excavating the intercorrelation between fully and weakly-supervised learning to improve the performance of object detection in RSIs. Initially, weakly-supervised learning (WSL) assigns the pseudo instancelevel annotations for the high-scoring positive bags to model the detector, which can be regarded as a weakened version of the region proposal network (RPN). Fortunately, RPN can be constrained by the ground truth of bounding boxes in fullysupervised learning (FSL), and the high-quality supervisions it provides are unavailable in any WSL methods. Moreover, we construct a proposal generation module (PGM), which further filters the unreliable bounding boxes, predicted by RPN, to supplement high-quality constraints into the ground truth and SS generated candidate boxes to supervise the optimization of WSL branch. By constructing an interactive learning paradigm of WSL and FSL, the former has more accurate constraints to learn an efficient auxiliary task, while the latter enjoys a richer representation form of data provided by WSL, which is undoubtedly a win-win process. Finally, we cascade the losses of WSL and FSL to further explore the intrinsic correlation between them by sharing the same feature extraction network. Experimental comparisons on DOTA and DIOR datasets demonstrate that our method achieves superior per
At present,most quantum secret sharing(QSS)protocols are more or less designed with the incorporation of classical secret sharing *** the increasing maturity of quantum technology,QSS protocols based on pure quantum m...
详细信息
At present,most quantum secret sharing(QSS)protocols are more or less designed with the incorporation of classical secret sharing *** the increasing maturity of quantum technology,QSS protocols based on pure quantum mechanics are becoming more *** secret sharing schemes cannot achieve absolute security,and their involvement can compromise the security of QSS *** paper proposes a QSS scheme based on Greenberger-Horn-Zeilinger(GHZ)basis measurement and quantum entanglement *** this protocol,the secret sender stores the secret information using Pauli *** obtain their shares by measuring the product state ***,participants complete the secret reconstruction through quantum entanglement exchange and other related quantum *** addition,the particles held by participants in the protocol do not contain any secret *** participant's particles are in a state of maximum entanglement,and no participant can deduce the particle information of other participants through their own *** the same time,the protocol is based on pure quantum mechanics and does not involve classical schemes,which avoids the problem of reduced security of the *** analysis indicates that the protocol is not vulnerable to retransmission interception and collusion ***,it is capable of detecting and terminating the protocol promptly when facing with attacks from dishonest participants.
The article addresses the output-feedback control issue for a class of multi-input multi-output(MIMO)uncertain nonlinear systems with multiple event-triggered mechanisms(ETM).Compared to previous event-triggering stud...
详细信息
The article addresses the output-feedback control issue for a class of multi-input multi-output(MIMO)uncertain nonlinear systems with multiple event-triggered mechanisms(ETM).Compared to previous event-triggering studies,this paper aims to trigger both the output and filtered *** nonlinear dynamics are approximated using fuzzy logic systems(FLSs).Then,a novel kind of state observer has been designed to deal with unmeasurable state problems using the triggered output *** sampled estimated state,the triggered output signal,and the filtered signal are utilized to propose an event-triggering mechanism that consists of sensor-to-observer(SO)and observer-to-controller(OC).An event-triggered output feedback control approach is given inside backstepping control,whereby the filter may be employed to circumvent the issue of the virtual control function not being differentiable at the trigger *** is testified that,according to the Lyapunov stability analysis scheme,all closed-loop signals and the system output are ultimately uniformly constrained by our control ***,the simulation examples are performed to confirm the theoretical findings.
Evolutionary algorithms have been extensively utilized in practical ***,manually designed population updating formulas are inherently prone to the subjective influence of the *** programming(GP),characterized by its t...
详细信息
Evolutionary algorithms have been extensively utilized in practical ***,manually designed population updating formulas are inherently prone to the subjective influence of the *** programming(GP),characterized by its tree-based solution structure,is a widely adopted technique for optimizing the structure of mathematical models tailored to real-world *** paper introduces a GP-based framework(GPEAs)for the autonomous generation of update formulas,aiming to reduce human *** modifications to tree-based GP have been instigated,encompassing adjustments to its initialization process and fundamental update operations such as crossover and mutation within the *** designing suitable function sets and terminal sets tailored to the selected evolutionary algorithm,and ultimately derive an improved update *** Cat Swarm Optimization Algorithm(CSO)is chosen as a case study,and the GP-EAs is employed to regenerate the speed update formulas of the *** validate the feasibility of the GP-EAs,the comprehensive performance of the enhanced algorithm(GP-CSO)was evaluated on the CEC2017 benchmark ***,GP-CSO is applied to deduce suitable embedding factors,thereby improving the robustness of the digital watermarking *** experimental results indicate that the update formulas generated through training with GP-EAs possess excellent performance scalability and practical application proficiency.
High reliability applications in dense access scenarios have become one of the main goals of 6G *** solve the access collision of dense Machine Type Communication(MTC)devices in cell-free communication systems,an inte...
详细信息
High reliability applications in dense access scenarios have become one of the main goals of 6G *** solve the access collision of dense Machine Type Communication(MTC)devices in cell-free communication systems,an intelligent cooperative secure access scheme based on multi-agent reinforcement learning and federated learning is proposed,that is,the Preamble Slice Orderly Queue Access(PSOQA)*** this scheme,the preamble arrangement is combined with the access *** preamble arrangement is realized by preamble slices which is from the virtual preamble *** access devices learn to queue orderly by deep reinforcement *** orderly queue weakens the random and avoids collision.A preamble slice is assigned to an orderly access queue at each access *** orderly queue is determined by interaction information among multiple *** the federated reinforcement learning framework,the PSOQA scheme is implemented to guarantee the privacy and security of ***,the access performance of PSOQA is compared with other random contention schemes in different load *** results show that PSOQA can not only improve the access success rate but also guarantee low-latency tolerant performances.
Spatial relations in text refer to how a geographical entity is located in space in relation to a reference *** spatial relations from text is a fundamental task in natural language *** studies have only focused on ge...
详细信息
Spatial relations in text refer to how a geographical entity is located in space in relation to a reference *** spatial relations from text is a fundamental task in natural language *** studies have only focused on generic fine-tuning methods with additional classifiers,ignoring the importance of the semantic correlation between different spatial elements and the large offset between the relation extraction task and the pre-trained *** address the above two issues,we propose a spatial relation extraction model based on Dual-view Prompt and Element Correlation(DPEC).Specifically,we first reformulate spatial relation extraction as a mask language model with a Dual-view Prompt(i.e.,Link Prompt and Confidence Prompt).Link Prompt can not only guide the model to incorporate more contextual information related to the spatial relation extraction task,but also better adapt to the original pre-training task of the language ***,Confidence Prompt can measure the confidence of candidate triplets in Link Prompt and work as a supplement to identify those easily confused examples in Link ***,we incorporate the element correlation to measure the consistency between different spatial elements,which is an effective cue for identifying the rationality of spatial *** results on the popular SpaceEval show that our DPEC significantly outperforms the SOTA baselines.
Two long-term slow slip events(SSEs) in Lower Cook Inlet, Alaska, were identified by Li SS et al.(2016). The earlier SSE lasted at least 9 years with M_(w) ~7.8 and had an average slip rate of ~82 mm/year. The latter ...
详细信息
Two long-term slow slip events(SSEs) in Lower Cook Inlet, Alaska, were identified by Li SS et al.(2016). The earlier SSE lasted at least 9 years with M_(w) ~7.8 and had an average slip rate of ~82 mm/year. The latter SSE, occurring in a similar area, lasted approximately 2 years with M_(w) ~7.2 and an average slip rate of ~91 mm/year. To test whether these SSEs triggered earthquakes near the slow slip area, we calculated the Coulomb stressing rate changes on receiver faults by using two fault geometry definitions: nodal planes of focal mechanism solutions of past earthquakes, and optimally oriented fault planes. Regions in the shallow slab(30–60 km) that experienced a significant increase in the Coulomb stressing rate due to slip by the SSEs showed an increase in seismicity rates during SSE periods. No correlation was found in the volumes that underwent a significant increase in the Coulomb stressing rate during the SSE within the crust and the intermediate slab. We modeled variations in seismicity rates by using a combination of the Coulomb stress transfer model and the framework of rate-and-state friction. Our model indicated that the SSEs increased the Coulomb stress changes on adjacent faults,thereby increasing the seismicity rates even though the ratio of the SSE stressing rate to the background stressing rate was small. Each long-term SSE in Alaska brought the megathrust updip of the SSE areas closer to failure by up to 0.1–0.15 MPa. The volumes of significant Coulomb stress changes caused by the Upper and Lower Cook Inlet SSEs did not overlap.
Cheng-Hu-Moruz(2017) completely classified the locally strongly convex centroaffine hypersurfaces with parallel cubic form based on the Calabi product(called the type I Calabi product for short) proposed by Li-Wan...
详细信息
Cheng-Hu-Moruz(2017) completely classified the locally strongly convex centroaffine hypersurfaces with parallel cubic form based on the Calabi product(called the type I Calabi product for short) proposed by Li-Wang(1991).In the present paper, the authors introduce the type II Calabi product(in case λ1 =2λ2), complementing the type I Calabi product(in case λ1 =≠ 2λ2), and achieve a classification of the locally strongly convex centroaffine hypersurfaces in Rn+1with vanishing centroaffine shape operator and Weyl curvature tensor by virtue of the types I and II Calabi *** a corollary, 3-dimensional complete locally strongly convex centroaffine hypersurfaces with vanishing centroaffine shape operator are completely classified, which positively answers the centroaffine Bernstein problems III and V by Li-Li-Simon(2004).
Smart contract security is crucial for blockchain applications. While studies suggest that only a small fraction of reported vulnerabilities are exploited, no follow-up research has investigated the reasons behind thi...
详细信息
暂无评论