Natural Language Inference (NLI) task requires an agent to determine the logical relationship between a natural language premise and hypothesis. We introduce Interactive Network (IIN), novel class of neural network architectures that is able achieve high-level understanding sentence pair by hierarchically extracting semantic features from interaction space. show tensor (attention weight) contains information solve inference, denser richer information. One instance such architecture, Densely...

Crop mapping is vital in ensuring food production security and informing governmental decision-making. The satellite-normalized difference vegetation index (NDVI) obtained during periods of vigorous crop growth important for species identification. Sentinel-2 images with spatial resolutions 10, 20, 60 m are widely used mapping. However, the often covered by clouds. In contrast, time-series moderate-resolution imaging spectrometer (MODIS) can usually capture phenology but coarse resolution....

An exact solution is given for the stress field due to an edge dislocation embedded in a three-phase composite cylinder. The force on then derived, from which set of simple approximate formulae also suggested. It shown that, comparison with two-phase model adopted by Dundurs and Mura (1964), allows have stable equilibrium position under much less stringent combinations material constants. As result, so-called trapping mechanism dislocations more likely take place model. Also, analysis...

Drone-view object detection (DroneDet) models typically suffer a significant performance drop when applied to nighttime scenes. Existing solutions attempt employ an exposure-adjustment module reveal objects hidden in dark regions before detection. However, most are only optimized for human perception, where the exposure-adjusted images may not necessarily enhance recognition. To tackle this issue, we propose novel Detection-driven Exposure-correction network DroneDet, called DEDet. The DEDet...

Super-resolution (SR) image reconstruction is a technique used to recover high-resolution using the cumulative information provided by several low-resolution images. With help of SR techniques, satellite remotely sensed images can be combined achieve higher-resolution image, which especially useful for two- or three-line camera satellite, e.g., ZY-3 Three Line Camera (TLC) satellite. In this paper, we introduce application method, including motion estimation and robust super-resolution...

Heat island phenomenon and air quality deterioration issues are two major problems that have occurred during the process of urbanization, especially in developing countries. A number measures been proposed, among which roof greening is considered as a promising one due to its outstanding performance thermal effects well improvement. self-maintenance system, termed Green Roof Manager (GRM), comprises irrigation shadowing subsystems, proposed this paper, focusing on automatic reliable...

The rapid development and increasing availability of high-resolution satellite (HRS) images provides increased opportunities to monitor large scale land cover. However, inefficiency excessive independence on expert knowledge limits the usage HRS a scale. As organization representation method, ontology can assist in improving efficiency automatic or semi-automatic cover information extraction, especially for images. This paper presents an ontology-based framework that was used model...

Since learning is typically very slow in Boltzmann machines, there a need to restrict connections within hidden layers. However, the resulting states of units exhibit statistical dependencies. Based on this observation, we propose using $l_1/l_2$ regularization upon activation possibilities restricted machines capture loacal dependencies among units. This not only encourages many groups be inactive given observed data but also makes group compete with each other for modeling data. Thus, RBMs...

Since learning in Boltzmann machines is typically quite slow, there a need to restrict connections within hidden layers. However, theresulting states of units exhibit statistical dependencies. Based on this observation, we propose using l1/l2 regularization upon the activation probabilities restricted capture local dependencies among units. This not only encourages many groups be inactive given observed data but also makes group compete with each other for modeling data. Thus, RBMs yields...

Image quality is important for taking full advantage of satellite data. As a common indicator, the National Imagery Interpretability Scale (NIIRS) widely used image assessment and provides comprehensive representation from perspective interpretability. The ZY-3 (Ziyuan-3) first civil high resolution mapping in China, which was established 2012. So far, there has been no reports on adopting NIIRS as indicator that This lack results gap between data users around world those China regarding...

Urban modeling and visualization are highly useful in the development of smart cities. Buildings most prominent features urban environment, necessary for decision support; thus, buildings should be modeled effectively efficiently three dimensions (3D). In this study, with help Gaofen-7 (GF-7) high-resolution stereo mapping satellite double-line camera (DLC) images multispectral (MUX) images, boundary a building is segmented via multilevel fusion network (MFFN). A digital surface model (DSM)...

Rapid urban development has led to severe land subsidence in Luoyang, Henan Province the past few years. The current research on areas of Luoyang is relatively and early, cannot reflect status Luoyang. To grasp prevent harm caused by subsidence. We used 57 Sentinel-1A Synthetic Aperture Radar (SAR) images covering obtain from April 2017 December 2021 using small baseline subset interferometric SAR technique. results show that there are several City, most serious area located near highway...

We apply the spike-and-slab Restricted Boltzmann Machine (ssRBM) to texture modeling. The ssRBM with tiled-convolution weight sharing (TssRBM) achieves or surpasses state-of-the-art on synthesis and inpainting by parametric models. also develop a novel RBM model visible layer binary variables in hidden layer. This is designed be stacked top of TssRBM. show resulting deep belief network (DBN) powerful generative that improves single-layer models capable modeling not only single...

The need for integrating geospatial information (GI) data from various heterogeneous sources has seen increased importance geographic system (GIS) interoperability. Using domain ontologies to clarify and integrate the semantics of is considered as a crucial step successful semantic integration in GI domain. Nevertheless, mechanisms are still needed facilitate mapping between described different natural languages. This research establishes formal ontology model cross-lingual mapping. By first...

For the purpose of evaluating image quality ZY-3 (Ziyuan-3) satellite imagery and from perspective data users in various professions, this paper proposed a user-oriented assessment to assess panchromatic multispectral images ZY-3, which included its statistical characteristics, visual effect, interpretability. In order better understand level product scope applications, comparison factors was conducted between SPOT-5 due close similarity their functionalities. As far as concerned, features...

Abstract GPS has a sharp performance decline in terms of accuracy indoors due to the complex building structure. A combined algorithm, targeting at received signal strength indication (RSSI) calibration optimisation, depending on deep neural network training via input vector Γ and target output Ψ, termed reference optimisation algorithm (RSOA) is proposed improve positioning indoor Bluetooth networks. Experimental results show that relative error RSOA between estimated measured ones can...

Medical image analysis, particularly polyp segmentation, plays a pivotal role in the early detection and prevention of colorectal cancer. The accuracy precision identification significantly influence subsequent diagnostic conclusions therapeutic strategies. However, segmentation still grapples with several challenges such as 1) considerable variations size, shape, color, location, 2) high degree visual similarity between polyps their immediate tissue surroundings, attributable to factors...