Search

Data augmentation with mixup has been proven effective in various machine learning tasks. However, previous methods primarily concentrate on generating previously unseen virtual examples using randomly selected mixed samples, which may overlook the importance of similar spatial distributions. In this work, we extend mixup and propose MbMix, a novel yet simple training approach designed for implementing mixup with memory batch augmentation. MbMix specifically selects the samples to be mixed via memory batch to guarantee that the generated samples have the same spatial distribution as the dataset samples. Conducting extensive experiments, we empirically validate that our method outperforms several mixup methods across a broad spectrum of text classification benchmarks, including sentiment classification, question type classification, and textual entailment. Of note, our proposed method achieves a 5.61% improvement compared to existing approaches on the TREC-fine benchmark. Our approach is versatile, with applications in sentiment analysis, question answering, and fake news detection, offering entrepreneurial teams and students avenues to innovate. It enables simulation and modeling for student ventures, fostering an entrepreneurial campus culture and mindset. [ABSTRACT FROM AUTHOR]

In this work, the Ag-SiO2-Ag metamaterial with elliptical nano-slits was proposed to investigate the multi-band polarization-dependent perfect absorber. It was found that multi-band perfect absorptions can be induced under TE and TM-polarized illuminations. Simulation results showed that the absorption peaks for TE-polarized wave appeared at 337.6 THz and 360.0 THz with 98.5% and 97.6% absorbance, respectively. Conversely, the absorption peaks for TM-polarized wave appeared at 325.7 THz and 366.1 THz with 96.3% and 97.9% absorbance, respectively. As a result, the elliptical metamaterial presented polarization-selectivity property for perfect absorption, and so, the metamaterial can filter out different frequencies of TE- and TM-reflected waves, i.e., the elliptical metamaterial can be used as a reflecting filter. In addition, this work studied the sensing performance of the elliptical metamaterial and showed that the dual-band sensing performances were different at low and high frequencies. The sensitivities (S) to the refractive index reached up to 151.1 THz/RIU and 120.8 THz/RIU for the TE and TM-polarized waves around 337.6 THz and 325.7 THz, which provide promising potential in near-infrared photoelectric sensor and detector. However, both the absorption frequency and intensity of TM-polarized wave were insensitive to the refractive index of the medium around 366.1 THz, and so, the study provides a theoretical basis for infrared stealth of different media. [ABSTRACT FROM AUTHOR]

Because of the high cost of data acquisition in synthetic aperture radar (SAR) target recognition, the application of synthetic (simulated) SAR data is becoming increasingly popular. Our study explores the problems encountered when training fully on synthetic data and testing on measured (real) data, and the distribution gap between synthetic and measured SAR data affects recognition performance under the circumstances. We propose a gradual domain adaptation recognition framework with pseudo-label denoising to solve this problem. As a warm-up, the feature alignment classification network is trained to learn the domain-invariant feature representation and obtain a relatively satisfactory recognition result. Then, we utilize the self-training method for further improvement. Some pseudo-labeled data are selected to fine-tune the network, narrowing the distribution difference between the training data and test data for each category. However, the pseudo-labels are inevitably noisy, and the wrong ones may deteriorate the classifier's performance during fine-tuning iterations. Thus, we conduct pseudo-label denoising to eliminate some noisy pseudo-labels and improve the trained classifier's robustness. We perform pseudo-label denoising based on the image similarity to keep the label consistent between the image and feature domains. We conduct extensive experiments on the newly published SAMPLE dataset, and we design two training scenarios to verify the proposed framework. For Training Scenario I, the framework matches the result of neural architecture searching and achieves 96.46% average accuracy. For Training Scenario II, the framework outperforms the results of other existing methods and achieves 97.36% average accuracy. These results illustrate the superiority of our framework, which can reach state-of-the-art recognition levels with appropriate stability. [ABSTRACT FROM AUTHOR]

Manual phenotyping of tomato plants is time consuming and labor intensive. Due to the lack of low-cost and open-access 3D phenotyping tools, the dynamic 3D growth of tomato plants during all growth stages has not been fully explored. In this study, based on the 3D structural data points generated by employing structures from motion algorithms on multiple-view images, we proposed a 3D phenotyping pipeline, 3DPhenoMVS, to calculate 17 phenotypic traits of tomato plants covering the whole life cycle. Among all the phenotypic traits, six of them were used for accuracy evaluation because the true values can be generated by manual measurements, and the results showed that the R2 values between the phenotypic traits and the manual ones ranged from 0.72 to 0.97. In addition, to investigate the environmental influence on tomato plant growth and yield in the greenhouse, eight tomato plants were chosen and phenotyped during seven growth stages according to different light intensities, temperatures, and humidities. The results showed that stronger light intensity and moderate temperature and humidity contribute to a higher biomass and higher yield. In conclusion, we developed a low-cost and open-access 3D phenotyping pipeline for tomato and other plants, and the generalization test was also complemented on other six species, which demonstrated that the proposed pipeline will benefit plant breeding, cultivation research, and functional genomics in the future. [ABSTRACT FROM AUTHOR]

The detection of ships on the open sea is an important issue for both military and civilian fields. As an active microwave imaging sensor, synthetic aperture radar (SAR) is a useful device in marine supervision. To extract small and weak ships precisely in the marine areas, polarimetric synthetic aperture radar (PolSAR) data have been used more and more widely. We propose a new PolSAR ship detection method which is based on a keypoint detector, referred to as a PolSAR-SIFT keypoint detector, and a patch variation indicator in this paper. The PolSAR-SIFT keypoint detector proposed in this paper is inspired by the SAR-SIFT keypoint detector. We improve the gradient definition in the SAR-SIFT keypoint detector to adapt to the properties of PolSAR data by defining a new gradient based on the distance measurement of polarimetric covariance matrices. We present the application of PolSAR-SIFT keypoint detector to the detection of ship targets in PolSAR data by combining the PolSAR-SIFT keypoint detector with the patch variation indicator we proposed before. The keypoints extracted by the PolSAR-SIFT keypoint detector are usually located in regions with corner structures, which are likely to be ship regions. Then, the patch variation indicator is used to characterize the context information of the extracted keypoints, and the keypoints located on the sea area are filtered out by setting a constant false alarm rate threshold for the patch variation indicator. Finally, a patch centered on each filtered keypoint is selected. Then, the detection statistics in the patch are calculated. The detection statistics are binarized according to the local threshold set by the detection statistic value of the keypoint to complete the ship detection. Experiments on three data sets obtained from the RADARSAT-2 and AIRSAR quad-polarization data demonstrate that the proposed detector is effective for ship detection. [ABSTRACT FROM AUTHOR]

Synthetic aperture radar (SAR) image change detection (CD) aims to automatically recognize changes over the same geographic region by comparing prechange and postchange SAR images. However, the detection performance is usually subject to several restrictions and problems, including the absence of labeled SAR samples, inherent multiplicative speckle noise, and class imbalance. More importantly, for bitemporal SAR images, changed regions tend to present highly variable sizes, irregular shapes, and different textures, typically referred to as hybrid variabilities, further bringing great difficulties to CD. In this paper, we argue that these internal hybrid variabilities can also be used for learning stronger feature representation, and we propose a hybrid variability aware network (HVANet) for completely unsupervised label-free SAR image CD by taking inspiration from recent developments in deep self-supervised learning. First, since different changed regions may exhibit hybrid variabilities, it is necessary to enrich distinguishable information within the input features. To this end, in shallow feature extraction, we generalize the traditional spatial patch (SP) feature to allow for each pixel in bitemporal images to be represented at diverse scales and resolutions, called extended SP (ESP). Second, with the carefully customized ESP features, HVANet performs local spatial structure information extraction and multiscale–multiresolution (MS-MR) information encoding simultaneously through a local spatial stream and a scale-resolution stream, respectively. Intrinsically, HVANet projects the ESP features into a new high-level feature space, where the change identification becomes easier. Third, to train the framework effectively, a self-supervision layer is attached to the top of the HVANet to enable the two-stream feature learning and recognition of changed pixels in the corresponding feature space, in a self-supervised manner. Experimental results on three low/medium-resolution SAR datasets demonstrate the effectiveness and superiority of the proposed framework in unsupervised SAR CD tasks. [ABSTRACT FROM AUTHOR]

Plant root system architecture changes drastically in response to availability of macronutrients in the soil environment. Despite the importance of root sulfur (S) uptake in plant growth and reproduction, molecular mechanisms underlying root development in response to S availability have not been fully characterized. We report here on the signaling module composed of the CLAVATA3 (CLV3)/EMBRYO SURROUNDING REGION (CLE) peptide and CLAVATA1 (CLV1) leucine-rich repeat receptor kinase, which regulate lateral root (LR) development in Arabidopsis thaliana upon changes in S availability. The wild-type seedlings exposed to prolonged S deficiency showed a phenotype with low LR density, which was restored upon sulfate supply. In contrast, the clv1 mutant showed a higher daily increase rate of LR density relative to the wild-type under prolonged S deficiency, which was diminished to the wild-type level upon sulfate supply, suggesting that CLV1 directs a signal to inhibit LR development under S-deficient conditions. CLE2 and CLE3 transcript levels decreased under S deficiency and through CLV1-mediated feedback regulations, suggesting the levels of CLE peptide signals are adjusted during the course of LR development. This study demonstrates a fine-tuned mechanism for LR development coordinately regulated by CLE-CLV1 signaling and in response to changes in S availability. [ABSTRACT FROM AUTHOR]

The development of advanced electrocatalysts for oxygen reduction and evolution is of paramount significance to fuel cells, water splitting, and metal-air batteries. Heteroatom-doped carbon materials have exhibited great promise because of their excellent electrical conductivity, abundance, and superior durability. Rationally optimizing active sites of doped carbons can remarkably enhance their electrocatalytic performance. In this study, nitrogen and oxygen codoped carbon nanotubes were readily synthesized from the pyrolysis of polydopamine-carbon nanotube hybrids. Different electron microscopes, Raman spectra and X-ray photoelectron spectroscopy (XPS) were employed to survey the morphological and componential properties. The newly-obtained catalyst features high-quality nitrogen and oxygen species, favourable porous structures and excellent electric conductivity, and thus exhibits remarkably bifunctional oxygen electrode activity. This research further helps to advance the knowledge of polydopamine and its potential applications as efficient electrocatalysts to replace noble metals. [ABSTRACT FROM AUTHOR]

This paper introduces a distributed secondary control algorithm for automatic generation control (AGC) and automatic voltage control (AVC), which aims at matching area generation to area load and minimizing the total generation cost in an alternating current (AC) microgrids. Firstly, the control algorithm utilizes a continuous-time distributed algorithm to generate additional control variables to achieve frequency-voltage recovery for all distributed generators (DGs). Secondary, it solves the economic dispatch problem (EDP) by a distributed economic incremental algorithm in the secondary control level, which avoids the problem caused by communication speed inconsistency between secondary and tertiary control levels. This study also utilizes a fully distributed strategy based on secondary communication network to estimate the total load demand. In addition, the proposed algorithm can be used to realize a seamless handover from the islanded mode to the grid-connected mode, run under the condition of short time communication system out of action, and help to realize the plug and play function. Lastly, the stability of the proposed control algorithm is analyzed and proved, and the effectiveness of the method is verified in some case studies. [ABSTRACT FROM AUTHOR]