共查询到20条相似文献,搜索用时 15 毫秒
1.
在传统的推荐算法中, 往往缺乏对用户长短期兴趣偏好问题的考虑, 而随着深度学习在推荐算法中应用的不断深入, 这一问题能够得到很好的解决. 本文针对该问题提出一种融合隐语义模型与门控循环单元的长短期推荐算法(recommendation algorithm based on long short-term, RA_LST), 以实现对用户长短期偏好的分别捕捉, 有效解决了因用户兴趣随时间变化而导致推荐效果下降的问题. 最终的实验结果表明, 本文提出的算法在不同的数据集上都表现出了推荐准确性的提升. 相似文献
2.
基于深度学习的自然环境下花朵识别 总被引:1,自引:0,他引:1
基于自然环境下的花朵识别已经成为了现在园艺植物以及计算机视觉方面的交叉研究热点。本文的花朵图像数据集是利用手机直接在自然场景中当场拍摄的,采集了湖南省植物园内26种观赏花朵的2600幅图像,其中还包括同一品种不同类别相似度很高的杜鹃,郁金香等花朵。设计了一种由3个残差块组成的20层深度学习模型Resnet20,模型的优化算法结合了Adam的高效初始化以及Sgd优秀的泛化能力,该优化算法主要是根据每次训练批次以及learning rate来进行转换调整,实验结果表明比单独使用Adam算法正确率高4到5个百分点,比单独使用Sgd算法收敛更快。该模型在Flower26数据集上,通过数据增强识别率可达到96.29%,表明深度学习是一种很有前途的应用于花朵识别的智能技术。 相似文献
3.
为提升真实场景视觉信号的采集质量,往往需要通过多种融合方式获取相应的图像,例如,多聚焦、多曝光、多光谱和多模态等。针对视觉信号采集的以上特性,图像融合技术旨在利用同一场景不同视觉信号的优势,生成单图像信息描述,提升视觉低、中、高级任务的性能。目前,依托端对端学习强大的特征提取、表征及重构能力,深度学习已成为图像融合研究的主流技术。与传统图像融合技术相比,基于深度学习的图像融合模型性能显著提高。随着深度学习研究的深入,一些新颖的理论和方法也促进了图像融合技术的发展,如生成对抗网络、注意力机制、视觉Transformer和感知损失函数等。为厘清基于深度学习技术的图像融合研究进展,本文首先介绍了图像融合问题建模,并从传统方法视角逐渐向深度学习视角过渡。具体地,从数据集生成、神经网络构造、损失函数设计、模型优化和性能评估等方面总结了基于深度学习的图像融合研究现状。此外,还讨论了选择性图像融合这类衍生问题建模(如基于高分辨率纹理图融合的深度图增强),回顾了一些基于图像融合实现其他视觉任务的代表性工作。最后,根据现有技术的缺陷,提出目前图像融合技术存在的挑战,并对未来发展趋势给出了展望。 相似文献
4.
图片问答是计算机视觉与自然语言处理交叉的多模态学习任务.为了解决该任务,研究人员提出堆叠注意力网络(stacked attention networks, SANs).研究发现该模型易陷入不好的局部最优解,引发较高的问答错误率.为了解决该问题,提出基于图片问答的静态重启随机梯度下降算法.实验结果和分析表明:它的准确率比基准算法提高0.29%,但其收敛速度慢于基准算法.为了验证改善性能的显著性,对实验结果进行统计假设检验.T检验结果证明它的改善性能是极其显著的.为了验证它在同类算法中的有效性,将该算法和当前最好的一阶优化算法进行有效性实验,实验结果和分析证明它更有效.为了验证它的泛化性能和推广价值,在经典的Cifar-10数据集上进行图像识别实验.实验结果和T检验结果证明:它具有良好的泛化性能和较好的推广价值. 相似文献
5.
针对现有隐含因子模型存在的新用户和项目的冷启动问题,提出基于用户分类的隐含因子模型,将用户分类信息融入到隐含因子的矩阵分解当中,先在原评分矩阵和用户分类信息的基础上使用指示函数和数据归一化等方法构建一个分类评分矩阵,再将分类评分矩阵融入到隐含因子模型的评分预测中。通过与传统隐含因子模型等方法在多个不同隐含因子个数上的实验比较分析,实验结果表明,改进模型能够不仅能解决新用户和项目的冷启动问题,还能有效降低预测评分的均方根误差,并提高预测推荐的准确度。 相似文献
6.
随机优化算法是求解大规模机器学习问题的高效方法之一.随机学习算法使用随机抽取的单个样本梯度代替全梯度,有效节省了计算量,但却会导致较大的方差.近期的研究结果表明:在光滑损失优化问题中使用减小方差策略,能够有效提高随机梯度算法的收敛速率.考虑求解非光滑损失问题随机优化算法COMID(compositeobjective mirror descent)的方差减小问题.首先证明了COMID具有方差形式的O(1/√T+σ2/√T)收敛速率,其中,T是迭代步数,σ2是方差.该收敛速率保证了减小方差的有效性,进而在COMID中引入减小方差的策略,得到一种随机优化算法α-MDVR(mirror descent with variance reduction).不同于Prox-SVRG(proximal stochastic variance reduced gradient),α-MDVR收敛速率不依赖于样本数目,每次迭代只使用部分样本来修正梯度.对比实验验证了α-MDVR既减小了方差,又节省了计算时间. 相似文献
7.
Mingsheng Shang Xin Luo Zhigang Liu Jia Chen Ye Yuan MengChu Zhou 《IEEE/CAA Journal of Automatica Sinica》2019,6(1):131-141
Latent factor (LF) models are highly effective in extracting useful knowledge from High-Dimensional and Sparse (HiDS) matrices which are commonly seen in various industrial applications. An LF model usually adopts iterative optimizers, which may consume many iterations to achieve a local optima, resulting in considerable time cost. Hence, determining how to accelerate the training process for LF models has become a significant issue. To address this, this work proposes a randomized latent factor (RLF) model. It incorporates the principle of randomized learning techniques from neural networks into the LF analysis of HiDS matrices, thereby greatly alleviating computational burden. It also extends a standard learning process for randomized neural networks in context of LF analysis to make the resulting model represent an HiDS matrix correctly. Experimental results on three HiDS matrices from industrial applications demonstrate that compared with state-of-the-art LF models, RLF is able to achieve significantly higher computational efficiency and comparable prediction accuracy for missing data. I provides an important alternative approach to LF analysis of HiDS matrices, which is especially desired for industrial applications demanding highly efficient models. 相似文献
8.
在机器学习领域中, 梯度下降算法是求解最优化问题最重要、最基础的方法. 随着数据规模的不断扩大, 传统的梯度下降算法已不能有效地解决大规模机器学习问题. 随机梯度下降算法在迭代过程中随机选择一个或几个样本的梯度来替代总体梯度, 以达到降低计算复杂度的目的. 近年来, 随机梯度下降算法已成为机器学习特别是深度学习研究的焦点. 随着对搜索方向和步长的不断探索, 涌现出随机梯度下降算法的众多改进版本, 本文对这些算法的主要研究进展进行了综述. 将随机梯度下降算法的改进策略大致分为动量、方差缩减、增量梯度和自适应学习率等四种. 其中, 前三种主要是校正梯度或搜索方向, 第四种对参数变量的不同分量自适应地设计步长. 着重介绍了各种策略下随机梯度下降算法的核心思想、原理, 探讨了不同算法之间的区别与联系. 将主要的随机梯度下降算法应用到逻辑回归和深度卷积神经网络等机器学习任务中, 并定量地比较了这些算法的实际性能. 文末总结了本文的主要研究工作, 并展望了随机梯度下降算法的未来发展方向. 相似文献
9.
Local SGD训练方法用于分布式机器学习以缓解通信瓶颈,但其本地多轮迭代特性使异构集群节点计算时间差距增大,带来较大同步时延与参数陈旧问题。针对上述问题,基于Local SGD方法提出了一种动态部分同步通信策略(LPSP),该方法利用两层决策充分发挥Local SGD本地迭代优势。在节点每轮迭代计算结束后,基于本地训练情况判断通信可能性,并在全局划分同步集合以最小化同步等待时延,减少Local SGD通信开销并有效控制straggler负面影响。实验表明LPSP可以在不损失训练精确度的情况下实现最高0.75~1.26倍的加速,此外,最高还有5.14%的精确度提升,可以有效加速训练收敛。 相似文献
10.
Hao Wu Xin Luo MengChu Zhou Muhyaddin J.Rawa Khaled Sedraoui Aiiad Albeshri 《IEEE/CAA Journal of Automatica Sinica》2022,9(3):533-546
A large-scale dynamically weighted directed network(DWDN) involving numerous entities and massive dynamic interaction is an essential data source in many big-data-related applications, like in a terminal interaction pattern analysis system(TIPAS). It can be represented by a high-dimensional and incomplete(HDI) tensor whose entries are mostly unknown. Yet such an HDI tensor contains a wealth knowledge regarding various desired patterns like potential links in a DWDN. A latent factorization-of-ten... 相似文献
11.
Karthick Panneerselvam K. Mahesh V. L. Helen Josephine A. Ranjith Kumar 《计算机系统科学与工程》2023,45(2):1047-1061
Deep learning has reached many successes in Video Processing. Video has become a growing important part of our daily digital interactions. The advancement of better resolution content and the large volume offers serious challenges to the goal of receiving, distributing, compressing and revealing high-quality video content. In this paper we propose a novel Effective and Efficient video compression by the Deep Learning framework based on the flask, which creatively combines the Deep Learning Techniques on Convolutional Neural Networks (CNN) and Generative Adversarial Networks (GAN). The video compression method involves the layers are divided into different groups for data processing, using CNN to remove the duplicate frames, repeating the single image instead of the duplicate images by recognizing and detecting minute changes using GAN and recorded with Long Short-Term Memory (LSTM). Instead of the complete image, the small changes generated using GAN are substituted, which helps with frame-level compression. Pixel wise comparison is performed using K-nearest Neighbours (KNN) over the frame, clustered with K-means and Singular Value Decomposition (SVD) is applied for every frame in the video for all three colour channels [Red, Green, Blue] to decrease the dimension of the utility matrix [R, G, B] by extracting its latent factors. Video frames are packed with parameters with the aid of a codec and converted to video format and the results are compared with the original video. Repeated experiments on several videos with different sizes, duration, Frames per second (FPS), and quality results demonstrated a significant resampling rate. On normal, the outcome delivered had around a 10% deviation in quality and over half in size when contrasted, and the original video. 相似文献
12.
Masaharu Kato Tetsuo Kosaka Akinori Ito Shozo Makino 《通讯和计算机》2009,6(5):28-35
Probabilistic Latent Semantic Analysis (PLSA) is proven to be effective in the information retrieval and the speech recognition technique. In this paper, we modify the calculation procedure of estimation algorithm. It substantially reduces the memory requirements. And, parallelization approach enables making models in less time. Next, we examined data segmentation for PLSA adaptation. Most meetings have a number of topics. We divide the meeting automatically and fit PLSA models with them. The experiment showed recognition performance improvement. 相似文献
13.
14.
15.
目的 多光谱图像融合是遥感领域中的重要研究问题,变分模型方法和深度学习方法是目前的研究热点,但变分模型方法通常采用线性先验构建融合模型,难以描述自然场景复杂非线性关系,导致成像模型准确性较低,同时存在手动调参的难题;而主流深度学习方法将融合过程当做一个黑盒,忽视了真实物理成像机理,因此,现有融合方法的性能依然有待提升。为了解决上述问题,提出了一种基于可解译深度网络的多光谱图像融合方法。方法 首先构建深度学习先验描述融合图像与全色图像之间的关系,基于多光谱图像是融合图像下采样结果这一认知构建数据保真项,结合深度学习先验和数据保真项建立一种新的多光谱图像融合模型,提升融合模型准确性。采用近端梯度下降法对融合模型进行求解,进一步将求解步骤映射为具有明确物理成像机理的可解译深度网络架构。结果 分别在Gaofen-2和GeoEye-1遥感卫星仿真数据集,以及QuickBird遥感卫星真实数据集上进行了主客观对比实验。相对于经典方法,本文方法的主观视觉效果有了显著提升。在Gaofen-2和GeoEye-1遥感卫星仿真数据集,相对于性能第2的方法,本文方法的客观评价指标全局相对无量纲误差(relat... 相似文献
16.
Ye Yuan;Siyang Lu;Xin Luo 《IEEE/CAA Journal of Automatica Sinica》2025,12(6):1246-1259
A non-negative latent factor (NLF) model is able to be built efficiently via a single latent factor-dependent, non-negative and multiplicative update (SLF-NMU) algorithm for performing precise representation to high-dimensional and incomplete (HDI) matrix from many kinds of big-data-related applications. However, an SLF-NMU algorithm updates a latent factor relying on the current update increment only without considering past learning information, making a resultant model suffer from slow convergence. To address this issue, this study proposes a proportional integral (PI) controller-enhanced NLF (PI-NLF) model with two-fold ideas: 1) Designing an increment refinement (IR) mechanism, which formulates the current and past update increments as the proportional and integral terms of a PI controller, thereby assimilating the past update information into the learning scheme smoothly with high efficiency; 2) Deriving an IR-based SLF-NMU (ISN) algorithm, which updates a latent factor following the principle of an IR mechanism, thus significantly accelerating an NLF model’s convergence rate. The simulation results on eight HDI matrices collected by real applications validate that a PI-NLF model outstrips several leading-edge models in both computational efficiency and accuracy when estimating missing data within an HDI matrix. The proposed PI-NLF model can be effectively applied to applications involving HDI matrix like e-commerce system, social network, and cloud service system. The code is available at https://github.com/yuanyeswu/PINLF/blob/main/PINLF-code.zip. 相似文献
17.
通过增加模型的深度以及训练数据的样本数量,深度神经网络模型能够在多个机器学习任务中获得更好的性能,然而这些必要的操作会使得深度神经网络模型训练的开销相应增大.因此为了更好地应对大量的训练开销,在分布式计算环境中对深度神经网络模型的训练过程进行加速成为了研发人员最常用的手段.随机梯度下降(stochastic gradient descent, SGD)算法是当前深度神经网络模型中最常见的训练算法之一,然而SGD在进行并行化的时候容易产生梯度过时问题,从而影响算法的整体收敛性.现有解决方案大部分针对的是各节点性能差别较小的高性能计算(high performance computing, HPC)环境,很少有研究考虑过各节点性能差别较大的集群环境.针对上述问题进行研究并提出了一种基于性能感知技术的动态batch size随机梯度下降算法(dynamic batch size SGD, DBS-SGD).该算法通过分析各节点的计算能力,对各节点的minibatch进行动态分配,从而保证了节点间每次迭代更新的时间基本一致,进而降低了节点的平均梯度过时值.提出的算法能够有效优化异步更新策略中存在的梯度过时问题.选用常用的图像分类基准Mnist和cifar10作为训练数据集,将该算法与异步随机梯度下降(asynchronous SGD, ASGD)算法、n-soft算法进行了对比.实验结果表明:在不损失加速比的情况下,Mnist数据集的loss函数值降低了60%,cifar数据集的准确率提升了约10%,loss函数值降低了10%,其性能高于ASGD算法和n-soft算法,接近同步策略下的收敛曲线. 相似文献
18.
Recommender systems have become a core part of the retail experience. Retailers often rely on recommender systems to help them drive more conversions through targeted communication and advertisements. However, recommender systems are not one size fits all. Specialized retailers require specialized recommender systems to consider various features, attributes, and dynamics about the product category. In this paper, we have proposed a novel fruit recommender system that generates dynamic recommendations while remediating the problem of data sparsity. We have developed a novel fruit recommender system that considers the temporal dynamics in the fruit market, like price fluctuations, fruit seasonality, and quality variations that occur throughout the year. To perform this task, we have used Recurrent Recommender Network (RRN), which uses the deep learning method Long Short-Term Memory (LSTM) to implement the system model. To ensure that our work and results obtained are practical, we have worked in a real-world setting, by tying up with a specialty fruit retailer based in New Delhi to get the real-world Point-of-Sale (POS) data of consumers. The result of the study suggests our algorithm performs better than other benchmark algorithms along NDCG and RMSE metrics. 相似文献
19.
This paper addresses the problem of course (path) generation when a learner's available time is not enough to follow the complete course. We propose a method to recommend successful paths regarding a learner's available time and his/her knowledge background. Our recommender is an instance of long term goal recommender systems (LTRS). This method, after locating a target learner in a course graph, applies a depth‐first search algorithm to find all paths for the learner given a time limitation. In addition, our method estimates learning time and score for all paths. It also indicates the probability of error for the estimated time and score for each path. Finally, our method recommends a path that satisfies the learner's time restriction while maximizing expected learning score. In order to evaluate our proposals for time and score estimation, we used the mean absolute error and average MAE. We have evaluated time and score estimation methods, including one proposed in the literature, on two E‐learning datasets. 相似文献
20.
深度学习已经广泛应用到各个领域, 如计算机视觉和自然语言处理等, 并都取得了明显优于早期机器学习算法的效果. 在信息技术飞速发展的今天, 训练数据逐渐趋于大数据集, 深度神经网络不断趋于大型化, 导致训练越来越困难, 速度和精度都有待提升. 2013年, Ioffe等指出训练深度神经网络过程中存在一个严重问题: 中间协变量迁移(Internal covariate shift), 使网络训练过程对参数初值敏感、收敛速度变慢, 并提出了批归一化(Batch normalization, BN)方法, 以减少中间协变量迁移问题, 加快神经网络训练过程收敛速度. 目前很多网络都将BN作为一种加速网络训练的重要手段, 鉴于BN的应用价值, 本文系统综述了BN及其相关算法的研究进展. 首先对BN的原理进行了详细分析. BN虽然简单实用, 但也存在一些问题, 如依赖于小批量数据集的大小、训练和推理过程对数据处理方式不同等, 于是很多学者相继提出了BN的各种相关结构与算法, 本文对这些结构和算法的原理、优势和可以解决的主要问题进行了分析与归纳. 然后对BN在各个神经网络领域的应用方法进行了概括总结, 并且对其他常用于提升神经网络训练性能的手段进行了归纳. 最后进行了总结, 并对BN的未来研究方向进行了展望. 相似文献