With the widespread deployment of indoor positioning systems, an unprecedented scale of indoor trajectories is being produced. By considering the inherent uncertainties and the text information contained in such an in...
详细信息
With the widespread deployment of indoor positioning systems, an unprecedented scale of indoor trajectories is being produced. By considering the inherent uncertainties and the text information contained in such an indoor trajectory, a new definition named Indoor Uncertain Semantic Trajectory is defined in this paper. In this paper, we focus on a new primitive, yet quite essential query named Indoor Uncertain Semantic Trajectory Similarity Join (IUST-Join for short), which is to match all similar pairs of indoor uncertain semantic trajectories from two sets. IUST-Join targets a number of essential indoor applications. With these applications in mind, we provide a purposeful definition of an indoor uncertain semantic trajectory similarity metric named IUS. To process IUST-Join more efficiently, both an inverted index on indoor uncertain semantic trajectories named 3IST and the first acceleration strategy are proposed to form a filtering-and-verification framework, where most invalid pairs of indoor uncertain semantic trajectories are pruned at quite low computation cost. And based on this filtering-and-verification framework, we present a highly-efficient algorithm named Indoor Uncertain Semantic Trajectory Similarity Join Processing (USP for short). In addition, lots of novel and effective acceleration strategies are proposed and embedded in the USP algorithm. Thanks to these techniques, both the time complexity and the time overhead of the USP algorithm are further reduced. The results of extensive experiments demonstrate the superior performance of the proposed work.
As a deep learning network with an encoder-decoder architecture,UNet and its series of improved versions have been widely used in medical image segmentation with great ***,when used to segment targets in 3D medical im...
详细信息
As a deep learning network with an encoder-decoder architecture,UNet and its series of improved versions have been widely used in medical image segmentation with great ***,when used to segment targets in 3D medical images such as magnetic resonance imaging(MRI),computed tomography(CT),these models do not model the relevance of images in vertical space,resulting in poor accurate analysis of consecutive slices of the same *** the other hand,the large amount of detail lost during the encoding process makes these models incapable of segmenting small-scale tumor *** at the scene of small-scale target segmentation in 3D medical images,a fully new neural network model SUNet++is proposed on the basis of UNet and UNet++.SUNet++improves the existing models mainly in three aspects:1)the modeling strategy of slice superposition is used to thoroughly excavate the three dimensional information of the data;2)by adding an attention mechanism during the decoding process,small scale targets in the picture are retained and amplified;3)in the up-sampling process,the transposed convolution operation is used to further enhance the effect of the *** order to verify the effect of the model,we collected and produced a dataset of hyperintensity MRI liver-stage images containing over 400 cases of liver *** results on both public and proprietary datasets demonstrate the superiority of SUNet++in small-scale target segmentation of three-dimensional medical images.
Federated learning (FL) is a promising decentralized machine learning approach that enables multiple distributed clients to train a model jointly while keeping their data private. However, in real-world scenarios, the...
详细信息
Federated learning (FL) is a promising decentralized machine learning approach that enables multiple distributed clients to train a model jointly while keeping their data private. However, in real-world scenarios, the supervised training data stored in local clients inevitably suffer from imperfect annotations, resulting in subjective, inconsistent and biased labels. These noisy labels can harm the collaborative aggregation process of FL by inducing inconsistent decision boundaries. Unfortunately, few attempts have been made towards noise-tolerant federated learning, with most of them relying on the strategy of transmitting overhead messages to assist noisy labels detection and correction, which increases the communication burden as well as privacy risks. In this paper, we propose a simple yet effective method for noise-tolerant FL based on the well-established co-training framework. Our method leverages the inherent discrepancy in the learning ability of the local and global models in FL, which can be regarded as two complementary views. By iteratively exchanging samples with their high confident predictions, the two models “teach each other” to suppress the influence of noisy labels. The proposed scheme enjoys the benefit of overhead cost-free and can serve as a robust and efficient baseline for noise-tolerant federated learning. Experimental results demonstrate that our method outperforms existing approaches, highlighting the superiority of our method.
Voronoi diagrams on triangulated surfaces based on the geodesic metric play a key role in many applications of computer *** methods of constructing such Voronoi diagrams generally depended on having an exact geodesic ...
详细信息
Voronoi diagrams on triangulated surfaces based on the geodesic metric play a key role in many applications of computer *** methods of constructing such Voronoi diagrams generally depended on having an exact geodesic ***,exact geodesic computation is time-consuming and has high memory usage,limiting wider application of geodesic Voronoi diagrams(GVDs).In order to overcome this issue,instead of using exact methods,we reformulate a graph method based on Steiner point insertion,as an effective way to obtain geodesic ***,since a bisector comprises hyperbolic and line segments,we utilize Apollonius diagrams to encode complicated structures,enabling Voronoi diagrams to encode a medial-axis surface for a dense set of boundary *** on these strategies,we present an approximation algorithm for efficient Voronoi diagram construction on triangulated *** also suggest a measure for evaluating similarity of our results to the exact *** our GVD results are constructed using approximate geodesic distances,we can get GVD results similar to exact results by inserting Steiner points on triangle *** results on many 3D models indicate the improved speed and memory requirements compared to previous leading methods.
Although the Mg-air battery with high theoretical energy density is desirable for the energy supply of marine engineering equipment,its applications remain limited due to the low actual discharge volt-age and inferior...
详细信息
Although the Mg-air battery with high theoretical energy density is desirable for the energy supply of marine engineering equipment,its applications remain limited due to the low actual discharge volt-age and inferior Mg anode utilization *** addition to the microstructure of Mg alloy anodes,the properties of discharge product films are of great importance to the discharge ***,the discharge behaviors of Mg-Y-Zn alloys are first studied mainly from the perspective of film *** contrastive analysis,it is found that the sufficient Y3+produced during the discharge process can substitute Mg2+in Mg(OH)2 to introduce effective cation *** Mg-Y-Zn anode with pro-fuse cation vacancies in the product film shows a synergy of potential and efficiency,and this can be attributed to an increase in the migration pathway for Mg2+,reducing the diffusion over-potential caused by the protective product *** study is expected to provide a new strategy from the perspective of cation vacancy design of discharge film for developing high-performance Mg-air batteries.
In recent years, significant progress has been made in knowledge graph representation learning, which has shown promising results in knowledge computing applications such as relation extraction and knowledge reasoning...
详细信息
Deterministic databases are able to reduce coordination costs in a *** property has fostered a significant interest in the design of efficient deterministic concurrency control ***,the state-of-the-art deterministic c...
详细信息
Deterministic databases are able to reduce coordination costs in a *** property has fostered a significant interest in the design of efficient deterministic concurrency control ***,the state-of-the-art deterministic concurrency control protocol Aria has three ***,it is impractical to configure a suitable batch size when the read-write set is ***,Aria running in low-concurrency scenarios,e.g.,a single-thread scenario,suffers from the same conflicts as running in high-concurrency ***,the single-version schema brings write-after-write *** address these issues,we propose Gria,an efficient deterministic concurrency control *** has the following ***,the batch size of Gria is ***,Gria’s conflict probability in low-concurrency scenarios is lower than that in high-concurrency ***,Gria has no write-after-write conflicts by adopting a multi-version *** further reduce conflicts,we propose two optimizations:a reordering mechanism as well as a rechecking *** evaluation result on two popular benchmarks shows that Gria outperforms Aria by 13x.
Passive daytime radiative cooling(PDRC)exhibits efficient cooling performance to reduce global fossil energy consumption and has attracted enormous *** efficient PDRC technology for achieving daytime subambient coolin...
详细信息
Passive daytime radiative cooling(PDRC)exhibits efficient cooling performance to reduce global fossil energy consumption and has attracted enormous *** efficient PDRC technology for achieving daytime subambient cooling is supposed to exhibit ultra-high solar reflectivity and infrared emissivity within atmospheric windows,among which ultra-high solar reflectivity is the principal *** most PDRC technologies,spherical particles are selected as scatterers to achieve high solar reflectivity because the Mie scattering effect of spherical scatterers can provide a high scattering ***,the weak backward scattering of spherical scatterers remains a *** this study,several nonspherical scatterers with strong backward scattering are introduced and researched *** numerical analysis results indicate that pyramidal scatterers provide the strongest backward scattering and square/circular scatterers exhibit the highest scattering ***,by considering scattering efficiency and asymmetry factor comprehensively,a dimensionless evaluation parameter is proposed,which can be employed as a quantitative design criterion to achieve the optimal-object-oriented designing of superior *** work can provide an innovative strategy for designing efficient passive daytime radiative cooling materials.
Generative adversarial networks(GANs) have drawn enormous attention due to their simple yet efective training mechanism and superior image generation quality. With the ability to generate photorealistic high-resolutio...
详细信息
Generative adversarial networks(GANs) have drawn enormous attention due to their simple yet efective training mechanism and superior image generation quality. With the ability to generate photorealistic high-resolution(e.g., 1024 × 1024) images, recent GAN models have greatly narrowed the gaps between the generated images and the real ones. Therefore, many recent studies show emerging interest to take advantage of pre-trained GAN models by exploiting the well-disentangled latent space and the learned GAN priors. In this study, we briefly review recent progress on leveraging pre-trained large-scale GAN models from three aspects, i.e.,(1) the training of large-scale generative adversarial networks,(2) exploring and understanding the pre-trained GAN models, and(3) leveraging these models for subsequent tasks like image restoration and editing.
We propose a scheme for generating high-quality single-photon sources utilizing the conventional photon blockade(CPB)effect in a cavity optomagnonic system with Kerr *** realization of the CPB effect depends on both t...
详细信息
We propose a scheme for generating high-quality single-photon sources utilizing the conventional photon blockade(CPB)effect in a cavity optomagnonic system with Kerr *** realization of the CPB effect depends on both the Kerr nonlinearity and Kerr-like nonlinearity of the optical cavity,which is converted using magneto-optical *** CPB effect can be realized in a cavity optomagnonic system with weak magneto-optical coupling by modulating the strength of the Kerr ***,our scheme supports photon blockade in both the strong and weak Kerr nonlinear regimes,which broadens the range of experimental ***,we explored the parameter regimes where the CPB effect could not be achieved because of the combined effects of the magneto-optical coupling and Kerr *** also determined the optimal driving amplitude region for generating high-quality single-photon *** work not only provides a novel route for realizing the CPB effect but also establishes a versatile platform for producing single-photon sources with high purity and brightness.
暂无评论