This paper demonstrates the novel approach of sub-micron-thick InGaAs broadband photodetectors(PDs)designed for high-resolution imaging from the visible to short-wavelength infrared(SWIR)*** approaches encounter chall...
详细信息
This paper demonstrates the novel approach of sub-micron-thick InGaAs broadband photodetectors(PDs)designed for high-resolution imaging from the visible to short-wavelength infrared(SWIR)*** approaches encounter challenges such as low resolution and crosstalk issues caused by a thick absorption layer(AL).Therefore,we propose a guided-mode resonance(GMR)structure to enhance the quantum efficiency(QE)of the InGaAs PDs in the SWIR region with only sub-micron-thick *** TiOx/Au-based GMR structure compensates for the reduced AL thickness,achieving a remarkably high QE(>70%)from 400 to 1700 nm with only a 0.98μm AL InGaAs PD(defined as 1μm AL PD).This represents a reduction in thickness by at least 2.5 times compared to previous results while maintaining a high ***,the rapid transit time is highly expected to result in decreased electrical *** effectiveness of the GMR structure is evident in its ability to sustain QE even with a reduced AL thickness,simultaneously enhancing the transit *** breakthrough offers a viable solution for high-resolution and low-noise broadband image sensors.
Semi-supervised learning techniques utilize both labeled and unlabeled images to enhance classification performance in scenarios where labeled images are limited. However, challenges such as integrating unlabeled imag...
详细信息
Semi-supervised learning techniques utilize both labeled and unlabeled images to enhance classification performance in scenarios where labeled images are limited. However, challenges such as integrating unlabeled images with incorrect pseudo-labels, determining appropriate thresholds for the pseudo-labels, and label prediction fluctuations on low-confidence unlabeled images, hinder the effectiveness of existing methods. This research introduces a novel framework named Interpolation Consistency for Bad Generative Adversarial Networks (IC-BGAN) that utilizes a new loss function. The proposed model combines bad adversarial training, fusion techniques, and regularization to address the limitations of semi-supervised learning. IC-BGAN creates three types of image augmentations and label consistency regularization in interpolation of bad fake images, real and bad fake images, and unlabeled images. It demonstrates linear interpolation behavior, reducing fluctuations in predictions, improving stability, and facilitating the identification of decision boundaries in low-density areas. The regularization techniques boost the discriminative capability of the classifier and discriminator, and send a better signal to the bad generator. This improves the generalization and the generation of diverse inter-class fake images as support vectors with information near the true decision boundary, which helps to correct the pseudo-labeling of unlabeled images. The proposed approach achieves notable improvements in error rate from 2.87 to 1.47 on the Modified National Institute of Standards and Technology (MNIST) dataset, 3.59 to 3.13 on the Street View House Numbers (SVHN) dataset, and 12.13 to 9.59 on the Canadian Institute for Advanced Research, 10 classes (CIFAR-10) dataset using 1000 labeled training images. Additionally, it reduces the error rate from 22.11 to 18.40 on the CINIC-10 dataset when using 700 labeled images per class. The experiments demonstrate the IC-BGAN framework outp
Predictability is an essential challenge for autonomous vehicles(AVs)’*** neural networks have been widely deployed in the AV’s perception ***,it is still an open question on how to guarantee the perception predicta...
详细信息
Predictability is an essential challenge for autonomous vehicles(AVs)’*** neural networks have been widely deployed in the AV’s perception ***,it is still an open question on how to guarantee the perception predictability for AV because there are millions of deep neural networks(DNNs)model combinations and system configurations when deploying DNNs in *** paper proposes configurable predictability testbed(CPT),a configurable testbed for quantifying the predictability in AV’s perception *** provides flexible configurations of the perception pipeline on data,DNN models,fusion policy,scheduling policies,and predictability *** top of CPT,the researchers can profile and optimize the predictability issue caused by different application and system *** has been open-sourced at:https://***/Torreskai0722/CPT.
Dear Editor,This letter presents a novel process monitoring model based on ensemble structure analysis(ESA).The ESA model takes advantage of principal component analysis(PCA),locality preserving projections(LPP),and m...
详细信息
Dear Editor,This letter presents a novel process monitoring model based on ensemble structure analysis(ESA).The ESA model takes advantage of principal component analysis(PCA),locality preserving projections(LPP),and multi-manifold projections(MMP)models,and then combines the multiple solutions within an ensemble result through Bayesian *** the developed ESA model,different structure features of the given dataset are taken into account simultaneously,the suitability and reliability of the ESA-based monitoring model are then illustrated through ***:The requirement for ensuring safe operation and improving process efficiency has led to increased research activity in the field of process monitoring.
We consider a power system whose electric demand pertaining to freshwater production is high(high freshwater electric demand),as in the Middle East,and investigate the tradeoff of storing freshwater in tanks versus st...
详细信息
We consider a power system whose electric demand pertaining to freshwater production is high(high freshwater electric demand),as in the Middle East,and investigate the tradeoff of storing freshwater in tanks versus storing electricity in batteries at the day-ahead operation *** storing freshwater and storing electricity increase the actual electric demand at valley hours and decrease it at peak hours,which is generally beneficial in term of cost and ***,to what extent?We analyze this question considering three power systems with different generation-mix configurations,i.e.,a thermal-dominated mix,a renewable-dominated one,and a fully renewable *** generation-mix configurations are inspired by how power systems may evolve in different countries in the Middle *** production uncertainty is compactly modeled using chance *** draw conclusions on how both storage facilities(freshwater and electricity)complement each other to render an optimal operation of the power system.
I had the privilege and the pleasure to work closely with Stephen J. Pennycook for about twenty years, having a group of post-docs and Vanderbilt-University graduate students embedded in his electron microscopy group ...
详细信息
I had the privilege and the pleasure to work closely with Stephen J. Pennycook for about twenty years, having a group of post-docs and Vanderbilt-University graduate students embedded in his electron microscopy group at Oak Ridge National Laboratory, spending on average a day per week there. We combined atomic-resolution imaging of materials,electron-energy-loss spectroscopy, and density-functional-theory calculations to explore and elucidate diverse materials phenomena, often resolving long-standing issues. This paper is a personal perspective of that journey, highlighting a few examples to illustrate the power of combining theory and microscopy and closing with an assessment of future prospects.
This study investigates the application of deep learning,ensemble learning,metaheuristic optimization,and image processing techniques for detecting lung and colon cancers,aiming to enhance treatment efficacy and impro...
详细信息
This study investigates the application of deep learning,ensemble learning,metaheuristic optimization,and image processing techniques for detecting lung and colon cancers,aiming to enhance treatment efficacy and improve survival *** introduce a metaheuristic-driven two-stage ensemble deep learning model for efficient lung/colon cancer *** diagnosis of lung and colon cancers is attempted using several unique indicators by different versions of deep Convolutional Neural Networks(CNNs)in feature extraction and model constructions,and utilizing the power of various Machine Learning(ML)algorithms for final ***,we consider different scenarios consisting of two-class colon cancer,three-class lung cancer,and fiveclass combined lung/colon cancer to conduct feature extraction using four *** extracted features are then integrated to create a comprehensive feature *** the next step,the optimization of the feature selection is conducted using a metaheuristic algorithm based on the Electric Eel Foraging Optimization(EEFO).This optimized feature subset is subsequently employed in various ML algorithms to determine the most effective ones through a rigorous evaluation *** top-performing algorithms are refined using the High-Performance Filter(HPF)and integrated into an ensemble learning framework employing weighted *** findings indicate that the proposed ensemble learning model significantly surpasses existing methods in classification accuracy across all datasets,achieving accuracies of 99.85%for the two-class,98.70%for the three-class,and 98.96%for the five-class datasets.
Gradient compression is a promising approach to alleviating the communication bottleneck in data parallel deep neural network (DNN) training by significantly reducing the data volume of gradients for synchronization. ...
详细信息
Gradient compression is a promising approach to alleviating the communication bottleneck in data parallel deep neural network (DNN) training by significantly reducing the data volume of gradients for synchronization. While gradient compression is being actively adopted by the industry (e.g., Facebook and AWS), our study reveals that there are two critical but often overlooked challenges: 1) inefficient coordination between compression and communication during gradient synchronization incurs substantial overheads, and 2) developing, optimizing, and integrating gradient compression algorithms into DNN systems imposes heavy burdens on DNN practitioners, and ad-hoc compression implementations often yield surprisingly poor system performance. In this paper, we propose a compression-aware gradient synchronization architecture, CaSync, which relies on flexible composition of basic computing and communication primitives. It is general and compatible with any gradient compression algorithms and gradient synchronization strategies and enables high-performance computation-communication pipelining. We further introduce a gradient compression toolkit, CompLL, to enable efficient development and automated integration of on-GPU compression algorithms into DNN systems with little programming burden. Lastly, we build a compression-aware DNN training framework HiPress with CaSync and CompLL. HiPress is open-sourced and runs on mainstream DNN systems such as MXNet, TensorFlow, and PyTorch. Evaluation via a 16-node cluster with 128 NVIDIA V100 GPUs and a 100 Gbps network shows that HiPress improves the training speed over current compression-enabled systems (e.g., BytePS-onebit, Ring-DGC and PyTorch-PowerSGD) by 9.8%-69.5% across six popular DNN models. IEEE
This study proposes a malicious code detection model DTL-MD based on deep transfer learning, which aims to improve the detection accuracy of existing methods in complex malicious code and data scarcity. In the feature...
详细信息
Purpose: The difficulty of diagnosing several lung disorders, including atelectasis, cardiomegaly, lung cancer, and COVID-19, is a challenging problem and needs to be addressed. These conditions exhibit some symptoms ...
详细信息
Purpose: The difficulty of diagnosing several lung disorders, including atelectasis, cardiomegaly, lung cancer, and COVID-19, is a challenging problem and needs to be addressed. These conditions exhibit some symptoms and demand advanced medical imaging process, thorough clinical assessments, and innovative procedures for accurate diagnosis. The shortage of qualified radiologists further makes the problem more complex to deal with. COVID-19 in particular has resulted in a remarkable number of fatalities around the world. Children below the age of 5 and individuals over 65 are more likely to be affected by lung disorders. It is very hard to diagnose and manage COVID-19 absolutely, but it can be identified earlier by employing computer-aided diagnosis (CAD) technologies to make timely diagnosis. Currently, radiologists adopt technologies, which are driven by artificial intelligence. By using them, medical imaging data, such as chest X-rays and CT scans, can be investigated to identify patterns to diagnose the severity of the virus. This expedites the diagnostic process and enhances accuracy, facilitating more timely and precise medical interventions. The efficiency of artificial intelligence in processing large datasets can directly help healthcare professionals in making diagnosis quicker and more accurate. The objective of the work in this paper is to design and implement deep learning model classifiers, which will effectively categorize the patterns found in the X-rays and CT scans. Methods: Three techniques for categorization are exploited to propose an entirely new hybrid convolutional neural network (CNN) model in this context. MRI and CT image categorization in the first classification method employ Fully Connected (FC) layers. The weights are calculated and tuned for training the algorithm over multiple periods to deliver the maximum precision for classification. The most accurate MRI and CT image characteristics are studied, and deep learning model classifiers
暂无评论