检索结果-南通市图书馆

Impact of Autotuned Fully Connected Layers on Performance of Self-supervised Models for Image Classification

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Machine Intelligence Research 2024年第6期21卷 1201-1213页

作者： Jaydeep kishore Snehasis Mukherjee Department of Computer Science and Engineering Shiv Nadar UniversityDelhi NCR 201314India

With the recent advancements of deep learning-based methods in image classification,the requirement of a huge amount of training data is inevitable to avoid overfitting ***,supervised deep learning models require labelled datasets for *** such a huge amount of labelled data requires considerable human effort and *** this scenario,self-supervised models are becoming popular because of their ability to learn even from unlabelled ***,the efficient transfer of knowledge learned by self-supervised models into a target task,is an unsolved *** paper proposes a method for the efficient transfer of know-ledge learned by a self-supervised model,into a target *** such as the number of layers,the number of units in each layer,learning rate,and dropout are automatically tuned in these fully connected(FC)layers using a Bayesian optimization technique called the tree-structured parzen estimator(TPE)approach *** evaluate the performance of the proposed method,state-of-the-art self-supervised models such as SimClr and SWAV are used to extract the learned *** are carried out on the CIFAR-10,CIFAR-100,and Tiny ImageNet *** proposed method outperforms the baseline approach with margins of 2.97%,2.45%,and 0.91%for the CIFAR-100,Tiny ImageNet,and CIFAR-10 datasets,respectively.

关键词： Neural architecture search autotuning self-supervised learning hyperparameter optimization tree-structured parzen estimator(TPE)

Segregation characteristics of irregular binaries in gas solid fluidized beds——An ANN-approach

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Particuology 2008年第3期6卷 199-206页

作者： Abanti Sahoo Gopendra kishore Roy Department of Chemical Engineering National Institute of TechnologyRourkela769008 OrissaIndia

Binary mixtures of irregular materials of different particle sizes and/or particle densities are fluidized in a 15-cm diameter column with a perforated plate distributor. An attempt has been made in this work to determine the segregation characteristics of jetsam particles for both the homogeneous and heterogeneous binary mixtures in terms of segregation distance by correlating it to the various system parameters, viz. initial static bed height, height of a layer of particles above the bottom grid, superficial gas velocity and average particle size and/or particle densities of the mixture through the dimensional analysis. Correlation on the basis of Artificial Neural Network approach has also been developed with the above system parameters thereby authenticating the development of correlation by the former approach. The calculated values of the segregation distance obtained for both the homogeneous and heterogeneous binary mixtures from both the types of ftuidized beds （i.e. under the static bed condition and the ftuidized bed condition） have also been compared with each other.

关键词： Gas-solid fluidization Segregation distance Static bed condition Fluidized bed condition Irregular binaries and artificial neural network

A numerical study of violent sloshing problems with modified MPS method

在线全文

学校读者我要写书评

暂无评论

Journal of Hydrodynamics 2017年第4期29卷 659-667页

作者： Debadatta Jena kishore Chandra Biswal Department of Civil Engineering National Institute of Technology

A numerical study on violent liquid sloshing phenomenon in a partially filled rectangular container is carried out by using moving particle semi-implicit（MPS） method. The present study deals with the implementation of five modifications all together over the original MPS method. The modifications include improved source terms for pressure Poisson equation, special approximation technique for the representation of gradient differential operator, collective action of mixed free surface particle identification boundary conditions, effecting Neumann boundary condition on solving the PPE and fixing judiciously the parting distance among particles to prevent collision. The suitability of the kernel function used in the original MPS method along with these five modifications is investigated for violent sloshing problems. The present model ensures a good agreement between numerical results with the existing experimental observations. The model is successfully applied to a partially filled tank undergoing horizontal sinusoidal excitation to compute the sloshing wave amplitudes and pressure on tank walls. The assessment of dynamic behaviour manifested in terms of base shear, overturning moment and impact pressure load exerted on tank ceiling induced by violent sloshing motion using MPS method is not reported in the open literature and has been efficiently carried out in the present study.

关键词： Sloshing particle method impact pressure base shear overturning moment

Seismic evaluation of frequency influence and resonant behavior for lead rubber bearing isolated rigid rectangular liquid tanks

在线全文

学校读者我要写书评

暂无评论

Earthquake Engineering and Engineering Vibration 2024年第4期23卷 973-994页

作者： Jyoti Ranjan Barik kishore Chandra Biswal Department of Civil Engineering National Institute of Technology RourkelaOdisha 769008India

The seismic behavior of a partially filled rigid rectangular liquid tank is investigated under short-and longduration ground motions.A finite element model is developed to analyze the liquid domain by using four-noded quadrilateral *** competency of the model is verified with the available *** studies are conducted for the dynamic parameters of the base-isolated tank,using a lead rubber bearing to evaluate the optimum damping and time period of the *** application of base isolation has reduced the total and impulsive hydrodynamic components of pressure by 80 to 90 percent,and base shear by 15 to 95 percent,depending upon the frequency content and duration of the considered *** sloshing amplitude of the base-isolated tank is reduced by 18 to 94 percent for most of the short-duration earthquakes,while it is increased by 17 to 60 percent for the majority of the long-duration ***,resonance studies are carried out through a long-duration harmonic excitation to obtain the dynamic behavior of non-isolated and isolated tanks,using a nonlinear sloshing *** seismic responses of the base-isolated tank are obtained as higher when the excitation frequency matches the fundamental sloshing frequency rather than the isolator frequency.

关键词： lead rubber bearing frequency domain resonant inputs long-duration excitation

Battery Charging Method for Electric Vehicles:From Wired to On-Road Wireless Charging

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Electrical Engineering 2018年第4期4卷 1-15页

作者： kishore Naik Mude Solace Power Inc. 1118 Topsail road Mount PearlNewfoundland and LabradorA1N5E7Canada

Recent fossil fuel shortages and global warming related problems have caused a substantial shift from internal combustion engine vehicles towards *** paper explores the thorough review of battery charging infrastructure from wired connection to on-road wireless charging for an *** initial part of the paper deals with the wired charging and its power electronics *** later portion deals with the wireless charging where both static and On-Road types are ***,various aspects of wireless power transfer are also *** Market scenario and future growth prospects are reviewed and presented in last section of the paper.

关键词： Wired charging static WPT on-road WPT

Mixed culture of Chlorella sp. and wastewater wild algae for enhanced biomass and lipid accumulation in artificial wastewater medium

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Environmental Science & Engineering 2018年第4期12卷 141-156页

作者： kishore Gopalakrishnan Javad Roostaei Yongli Zhang Department of Civil and Environmental Engineering Wayne State University Detroit MI 48201 USA＊

The purpose of this work is to study the co-cultivation of Chlorella sp. and wastewater wild algae under different cultivation conditions （i.e. CO2, light intensity, cultivation time, and inoculation ratio） for enhanced algal biomass and lipid productivity in wastewater medium using Response SuHhce Methodology （RSM）. The results show that mixed cultures ofd7llorella sp. and wastewater wild algae increase biomass and lipid yield. Additionally, findings indicate that CO2, light intensity and cultivation time significantly affect algal productivity. Furthcnnore, CO2 concentration and light intensity, and CO2 concentration and algal composition, have an interactive effect on biomass productivity. Under dii＂ferent cultivation conditions, the response of algal biomass, cell count, and lipid productlvlty ranges from2,5 to 10.2 mg/mL 1.1 × 10 to 8.2 × 10 cells/mL and 1.1 × 10^6 to 6.8 × 10^12 total fluorescent units/mL, respectively× The optimum conditions tbr simt, ltaneot, s biomass and lipid accumulation are 3.6% of CO2 （v/v）, 160 μmol/m^2/s of light intensity, 1×6/2.4 of inoculation ratio （wastewater-algae/Chlorella）, and 8.3 days of cultivation time. The optimal productivity is 9,8 （g/L） for dry biomass, 8.6 E ＋ 08 （cells/mL） for cell count, and 6.8 E ＋ 12 （Total FL units per mL） fbr lipid yield, achieving up to four times, eight times, and seven times higher productivity compared to non- optimized conditions. Provided is a supportive methodology to improve mixed algal culture for hioenergy feedstock generation and to optimize cultivation conditions in complex wastewater environments. This work is an important step tbrward in the development of sustainable large-scale algae cultivation for cost-efficient generation of biofuel.

关键词： Algal biofuels Algal mixed cultures Algal biomass Algal lipid Wastewater Response surface methodology

Short-term mixed electricity demand and price forecasting using adaptive autoregressive moving average and functional link neural network

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Modern Power Systems and Clean Energy 2019年第5期7卷 1241-1255页

作者： Sujit Kumar DASH Pradipta kishore DASH Siksha O Anusandhan Deemed to be University Bhubaneswar 751030India

A new hybrid adaptive autoregressive moving average(ARMA)and functional link neural network(FLNN)trained by adaptive cubature Kalman filter(ACKF)is presented in this paper for forecasting day-ahead mixed short-term demand and electricity prices in smart *** hybrid forecasting framework is intended to capture the dynamic interaction between the electricity consumers and the forecasted prices resulting in the shift of demand curve in electricity *** proposed model comprises a linear ARMA-FLNN obtained by using a nonlinear expansion of the weighted *** nonlinear functional block helps introduce nonlinearity by expanding the input space to higher dimensional space through basis *** train the ARMA-FLNN,an ACKF is used to obtain faster convergence and higher forecasting *** proposed method is tested on several electricity markets,and the performance metrics such as the mean average percentage error(MAPE)and error variance are compared with other forecasting methods,indicating the improved accuracy of the approach and its suitability for a real-time forecasting.

关键词： Mixed demand and price forecasting Cubature Kalman filter Price spikes Smart grids

A novel control strategy of a variable-speed doubly-fed-induction-generator-based wind energy conversion system

在线全文

学校读者我要写书评

暂无评论

Clean Energy 2024年第1期8卷 153-170页

作者： Shibani Prasad Mohapatra Pradipta kishore Dash Department of Electrical Engineering Siksha O Anusandhan Deemed to be UniversityBhubaneswar-751030OdishaIndia Multidisciplinary Research Cell Siksha O Anusandhan Deemed to be UniversityBhubaneswar-751030OdishaIndia

This paper aims to address the issue of control of a variable-speed wind turbine based on doubly-fed induction generators. In this work,an effort is made to extract the maximum efficiency from a doubly-fed induction generator-based variable-speed wind turbine by controlling the rotor current. In the first step, a maximum power point tracking technique is used to extract the maximum power from theturbine. Then a stator-flux-oriented vector control strategy is employed to control the rotor-side current. Subsequently, a grid voltagevector-oriented control strategy is used to control the grid-side system of the grid-connected generator. Considering the nonlinearityand parameter uncertainty of the system, an active disturbance rejection controller with a sliding-mode-based extended-state observeris developed for the above-mentioned control strategies. Furthermore, the stability of the controller is tested and the performance of thecontroller is compared with the classical proportional-integral controller based on disturbance rejection, robustness and tracking capability in a highly non-linear wind speed variation scenario. Modelling, control and comparison are conducted in the MATLAB®/Simulink®environment. Finally, a real-time hardware set-up is presented using the dSPACE ds-1104 R&D processing board to validate the controlscheme. From the result of the experiments, it is seen that the proposed controller takes 10-15 control cycles to settle to its steady-statevalues, depending on the control loop, whereas the conventional proportional-integral controller takes 60-75 control cycles. As a result,the settling time for the proposed control scheme is shorter than that of the proportional-integral controller.

关键词： doubly-fed induction generator variable-speed wind turbine vector control active disturbance rejection controller stability analysis

Human and Machine Vision Based Indian Race Classification Using Modified-Convolutional Neural Network

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computer Systems Science & Engineering 2023年第3期44卷 2603-2618页

作者： Vani A.Hiremani kishore Kumar Senapati Department of Computer Science&Engineering Birla Institute of TechnologyMesraRanchi835215JharkhandIndia

The inter-class face classification problem is more reasonable than the intra-class classification *** address this issue,we have carried out empirical research on classifying Indian people to their geographical *** work aimed to construct a computational classification model for classifying Indian regional face images acquired from south and east regions of India,referring to human *** have created an Automated Human Intelligence System(AHIS)to evaluate human visual *** of AHIS response showed that face shape is a discriminative feature among the other facial *** have developed a modified convolutional neural network to characterize the human vision response to improve face classification *** proposed model achieved mean F1 and Matthew Correlation Coefficient(MCC)of 0.92 and 0.84,respectively,on the validation set,outperforming the traditional Convolutional Neural Network(CNN).The CNN-Contoured Face(CNN-FC)model is developed to train contoured face images to investigate the influence of face ***,to cross-validate the accuracy of these models,the traditional CNN model is trained on the same *** an accuracy of 92.98%,the Modified-CNN(M-CNN)model has demonstrated that the proposed method could facilitate the tangible impact in intra-classification problems.A novel Indian regional face dataset is created for supporting this supervised classification work,and it will be available to the research community.

关键词： Data collection and preparation human vision analysis machine vision canny edge approximation method color local binary patterns convolutional neural network