基于PolyLoss的StarGAN图像样本自生长研究

在深度学习模型训练中,提高输入样本数量能够有效解决样本不均衡和训练过拟合问题。为更好地处理数据集的小样本场景,提出了一种基于PolyLoss的StarGAN小样本图像增强方法（PolyStarGAN）。相比传统样本增强方法,该网络具有更高的样本学习、生成能力。该方法使用训练生成器和判别器提取图像的特征信息,生成图像与输入图像的内容和结构一致。通过计算峰值信噪比、结构相似性与视觉信息保真度3个方面评估PolyStarGAN网络生成结果;在ResNet50目标识别模型训练中验证数据集质量。实验结果表明,提出的模型对不同场景图像的增强效果较好,提高了自生长数据集的ResNet50模型识别精度,可为样本自生长中的小样本问题提供参考方案与技术支持。     

基于栈式去噪自动编码器的边际Fisher分析算法

特征学习是模式识别领域的关键问题。基于自动编码器的深度神经网络通过无监督预训练与有监督微调能够有效地提取数据中关键信息,形成特征。提出一种基于栈式去噪自编码器的边际Fisher分析算法,该算法将边际Fisher分析运用于有监督微调阶段,进一步提升算法的特征学习能力。实验结果表明,该算法与标准的栈式去噪自编码器和基于受限玻尔兹曼机的深度信念网相比,具有更好的识别效果。     

基于信息融合的模式识别方法研究

将多特征多分类器的模式识别看作信息融合问题,提出了一种基于支持向量机理论的决策级信息融合模型,并讨论了基于该模型的识别精度以及学习样本数之间的关系.通过对ORL人脸图像库、CENPAMI手写体数字数据库这两类不同样本问题的识别实验表明,基于信息融合的模式识别方法可以取得较高的整体识别性能.     

基于块稀疏贝叶斯学习的人体运动模式识别

在人体运动模式识别中, 传统稀疏表示分类算法未考虑待测试样本相应稀疏系数向量内在块结构相关性信息,影响了算法识别性能。为此,提出一种基于块稀疏模型的人体运动模式识别方法。该方法充分利用人体运动模式内在块稀疏结构,将人体运动模式识别问题转化为稀疏表示问题,采用块稀疏贝叶斯学习算法,求解基于样本训练集优化稀疏表示待测样本的稀疏系数, 并根据稀疏系数重构残差判定待识别动作类别,能有效提高人体运动模式识别率。选用包含多类别人体动作行为模式的USC-HAD数据库对所提算法性能进行了验证。实验结果表明,所提算法能够有效捕获不同运动模式内在差异信息,平均动作识别率达到97.86%,比传统动作识别方法平均提高近5%,有效提高了动作识别准确率。     

双流网络信息交互机制下的微表情识别

针对深度学习的方法用于微表情识别时微表情识别的实验数据库非常稀缺,导致神经网络在学习的过程中知识获取有限而难以提高精度及泛化能力的问题,提出基于双流网络信息交互的微表情识别方法.通过改进的深度互学习策略引导图像序列不同模态之间的交互训练,提高网络的识别率.方法基于RGB图像序列建立主体网络,基于光流建立辅助网络;在训练阶段,通过设计互学习损失中的有监督学习损失和拟态损失,优化训练过程,使得每一种模态都能学习正确地预测训练样本的真实标识,同时能与其他模态的预测相匹配;在测试阶段,由于互学习机制增强了RGB分支的判别能力,因此可对光流分支进行剪裁,在保证精度的前提下提高识别速度.在CASME,CASMEⅡ和SMIC数据库上的实验结果表明,该方法有效地提高了识别精度,整体性能优于已有方法.     

支持向量学习机在点云去噪中的应用

在解决非线性、高维模式识别以及小样本等问题中,支持向量机表现出许多独有的优势.提出将支持向量机学习分类方法应用于点云去噪中,能够稳定地进行机器学习,训练得到判别模型,快速、准确地识别出噪声点与非噪声点.通过对小样本数据的统计学习,能够推广到大规模数据中去进行结果的预测估计.用SVM对点云数据样本进行学习训练、测试,识别分类,从而达到去噪光顺的目的.实验表明,此方法在有效去除噪声的同时能较完整地保留点云数据信息.     

基于u-wordMixup的半监督深度学习模型

当标注样本匮乏时,半监督学习利用大量未标注样本解决标注瓶颈的问题,但由于未标注样本和标注样本来自不同领域,可能造成未标注样本存在质量问题,使得模型的泛化能力变差,导致分类精度下降.为此,基于wordMixup方法,提出针对未标注样本进行数据增强的u-wordMixup方法,结合一致性训练框架和Mean Teacher模型,提出一种基于u-wordMixup的半监督深度学习模型(semi-supervised deep learning model based on u-wordMixup,SD-uwM).该模型利用u-wordMixup方法对未标注样本进行数据增强,在有监督交叉熵和无监督一致性损失的约束下,能够提高未标注样本质量,减少过度拟合.在AGNews、THUCNews和20 Newsgroups数据集上的对比实验结果表明,所提出方法能够提高模型的泛化能力,同时有效提高时间性能.     

改进的小脑控制神经网络算法研究

研究小脑控制器神经网络(CMAC)在模式识别中的应用问题.在算法优化过程中,针对模式识别中训练样本中存在着高维、大量冗余信息而传统CMAC不能够对输入信息空间维数降低,从而常导致CMAC网络训练速度慢、识别率低等问题.为了加快CMAC学习速度,提高识别率,提出一种基于粗糙集(RS)的CMAC模式识别方法(RS_CMAC).利用粗糙集约简并删除训练样本中的冗余信息,降低样本的维数,从而优化网络的结构;同时网络中引入了自适应动态学习率,加快网络的收敛速度和学习速度,从而提高识别率.以数码管模式识别为例对算法进行了验证性实验,仿真结果表明,与传统CMAC相比.识别精度有了明显提高,学习速度加快.RS_CMAC方法克服了传统CMAC缺点和不足,是一种有效的模式识别方法,为实际应用提供依据.     

基于多学习器协同训练模型的人体行为识别方法

人体行为识别是计算机视觉研究的热点问题,现有的行为识别方法都是基于监督学习框架.为了取得较好的识别效果,通常需要大量的有标记样本来建模.然而,获取有标记样本是一个费时又费力的工作.为了解决这个问题,对半监督学习中的协同训练算法进行改进,提出了一种基于多学习器协同训练模型的人体行为识别方法.这是一种基于半监督学习框架的识别算法.该方法首先通过基于Q统计量的学习器差异性度量选择算法来挑取出协同训练中基学习器集,在协同训练过程中,这些基学习器集对未标记样本进行标记;然后,采用了基于分类器成员委员会的标记近邻置信度计算公式来评估未标记样本的置信度,选取一定比例置信度较高的未标记样本加入到已标记的训练样本集并更新学习器来提升模型的泛化能力.为了评估算法的有效性,采用混合特征来表征人体行为,从而可以快速完成识别过程.实验结果表明,所提出的基于半监督学习的行为识别系统可以有效地辨识视频中的人体动作.     

基于单类分类器的半监督学习

提出一种结合单类学习器和集成学习优点的Ensemble one-class半监督学习算法.该算法首先为少量有标识数据中的两类数据分别建立两个单类分类器.然后用建立好的两个单类分类器共同对无标识样本进行识别,利用已识别的无标识样本对已建立的两个分类面进行调整、优化.最终被识别出来的无标识数据和有标识数据集合在一起训练一个基分类器,多个基分类器集成在一起对测试样本的测试结果进行投票.在5个UCI数据集上进行实验表明,该算法与tri-training算法相比平均识别精度提高4.5%,与仅采用纯有标识数据的单类分类器相比,平均识别精度提高8.9%.从实验结果可以看出,该算法在解决半监督问题上是有效的.     

A Hybrid Training Algorithm for Feedforward Neural Networks

Feedforward neural networks (FNN) have been proposed to solve complex problems in pattern recognition, classification and function approximation. Despite the general success of learning methods for FNN, such as the backpropagation (BP) algorithm, second-order algorithms, long learning time for convergence remains a problem to be overcome. In this paper, we propose a new hybrid algorithm for a FNN that combines unsupervised training for the hidden neurons (Kohonen algorithm) and supervised training for the output neurons (gradient descent method). Simulation results show the effectiveness of the proposed algorithm compared with other well-known learning methods.     

感知器在语言模型训练中的应用

感知器（perceptron）是神经网络模型中的一种,它可以通过监督学习（supervised learning）的方法建立模式识别的能力．将感知器应用到语言模型的训练中,实现了感知器的两种不同训练规则以及多种特征权值计算方法,讨论了不同的训练参数对训练效果的影响．在训练之前,使用了一种基于经验风险最小化（empirical risk minimization,ERM）的特征选择算法确定特征集合．感知器训练之后的语言模型在日文假名到汉字（kana-kanji）的转换中进行评估．通过实验对比了感知器的两种训练规则以及变形算法的性能,同时发现通过感知器训练的模型比传统模型（N-gram）在性能上有了很大的提高,使相对错误率下降了15％～20％．     

Adaptive nearest neighbor pattern classification

A variant of nearest-neighbor (NN) pattern classification and supervised learning by learning vector quantization (LVQ) is described. The decision surface mapping method (DSM) is a fast supervised learning algorithm and is a member of the LVQ family of algorithms. A relatively small number of prototypes are selected from a training set of correctly classified samples. The training set is then used to adapt these prototypes to map the decision surface separating the classes. This algorithm is compared with NN pattern classification, learning vector quantization, and a two-layer perceptron trained by error backpropagation. When the class boundaries are sharply defined (i.e., no classification error in the training set), the DSM algorithm outperforms these methods with respect to error rates, learning rates, and the number of prototypes required to describe class boundaries.     

一种新的特征提取方法及其在模式识别中的应用

核典型相关分析(KCCA)是一种有监督的机器学习方法,可以有效地提取非线性特征。然而随着训练样本数目的增加,标准的KCCA方法的计算复杂度会随之增加。针对此缺点,提出一种改进的KCCA方法：首先用几何特征选择方法选择一个训练样本子集并将其映射到再生核希尔伯特空间(RKHS),然后设计了一种提升特征提取效率的算法,该算法按照对特征分类贡献的大小巧妙地选取样本的特征值,进而求出其相应的特征向量,最后将改进的KCCA与支持向量数据描述(SVDD)多分类器相结合用于分类识别。在ORL人脸图像数据库上的实验结果表明,改进的方法相对传统的KCCA方法,在不影响识别率的情况下提高了人脸识别速度,减小了系统存储量。     

基于核聚类方法的多层次支持向量机分类树

针对解决多类模式识别问题的SVM方法进行研究。在比较几种常用的多类SVM分类算法的基础上，提出一种基于核聚类方法的多层次SVM分类树，将核空问中的无监督学习方法和有监督学习方法结合起来，实现了一种结构更加简洁清晰、计算效率更高的多层SVM分类树算法，并在实验中取得了良好的结果．     

ART-EMAP: A neural network architecture for object recognition byevidence accumulation

A new neural network architecture is introduced for the recognition of pattern classes after supervised and unsupervised learning. Applications include spatio-temporal image understanding and prediction and 3D object recognition from a series of ambiguous 2D views. The architecture, called ART-EMAP, achieves a synthesis of adaptive resonance theory (ART) and spatial and temporal evidence integration for dynamic predictive mapping (EMAP). ART-EMAP extends the capabilities of fuzzy ARTMAP in four incremental stages. Stage 1 introduces distributed pattern representation at a view category field. Stage 2 adds a decision criterion to the mapping between view and object categories, delaying identification of ambiguous objects when faced with a low confidence prediction. Stage 3 augments the system with a field where evidence accumulates in medium-term memory. Stage 4 adds an unsupervised learning process to fine-tune performance after the limited initial period of supervised network training. Each ART-EMAP stage is illustrated with a benchmark simulation example, using both noisy and noise-free data.     

Combined use of supervised and unsupervised learning for power system dynamic security mapping

This paper proposes a new methodology which combines supervised and unsupervised learning for evaluating power system dynamic security. Based on the concept of stability margin, pre-fault power system conditions are assigned to the output neurons on the two-dimensional grid with the growing hierarchical self-organizing map technique (GHSOM) via supervised artificial neural networks (ANNs) which perform an estimation of post-fault power system state. The technique estimates the dynamic stability index that corresponds to the most critical value of synchronizing and damping torques of multimachine power systems. ANN-based pattern recognition is carried out with the growing hierarchical self-organizing feature mapping in order to provide adaptive neural network architecture during its unsupervised training process. Numerical tests, carried out on a IEEE 9 bus power system are presented and discussed. The analysis using such method provides accurate results and improves the effectiveness of system security evaluation.     

SemiBoost: Boosting for Semi-Supervised Learning

Semi-supervised learning has attracted a significant amount of attention in pattern recognition and machine learning. Most previous studies have focused on designing special algorithms to effectively exploit the unlabeled data in conjunction with labeled data. Our goal is to improve the classification accuracy of any given supervised learning algorithm by using the available unlabeled examples. We call this as the Semi-supervised improvement problem, to distinguish the proposed approach from the existing approaches. We design a metasemi-supervised learning algorithm that wraps around the underlying supervised algorithm and improves its performance using unlabeled data. This problem is particularly important when we need to train a supervised learning algorithm with a limited number of labeled examples and a multitude of unlabeled examples. We present a boosting framework for semi-supervised learning, termed as SemiBoost. The key advantages of the proposed semi-supervised learning approach are: 1) performance improvement of any supervised learning algorithm with a multitude of unlabeled data, 2) efficient computation by the iterative boosting algorithm, and 3) exploiting both manifold and cluster assumption in training classification models. An empirical study on 16 different data sets and text categorization demonstrates that the proposed framework improves the performance of several commonly used supervised learning algorithms, given a large number of unlabeled examples. We also show that the performance of the proposed algorithm, SemiBoost, is comparable to the state-of-the-art semi-supervised learning algorithms.     

Bayesian sample steered discriminative regression for biometric image classification

Regression techniques, such as ridge regression (RR) and logistic regression (LR), have been widely used in supervised learning for pattern classification. However, these methods mainly exploit the class label information for linear mapping function learning. They will become less effective when the number of training samples per class is small. In visual classification tasks such as face recognition, the appearance of the training sample images also conveys important discriminative information. This paper proposes a novel regression based classification model, namely Bayesian sample steered discriminative regression (BSDR), which simultaneously exploits the sample class label and the sample appearance for linear mapping function learning by virtue of the Bayesian formula. BSDR learns a linear mapping for each class to extract the image class label features, and classification can be simply done by nearest neighbor classifier. The proposed BSDR method has advantages such as small number of mappings, insensitiveness to input feature dimensionality and robustness to small sample size. Extensive experiments on several biometric databases also demonstrate the promising classification performance of our method.     

A semi-supervised convolutional neural network-based method for steel surface defect recognition

Automatic defect recognition is one of the research hotspots in steel production, but most of the current methods focus on supervised learning, which relies on large-scale labeled samples. In some real-world cases, it is difficult to collect and label enough samples for model training, and this might impede the application of most current works. The semi-supervised learning, using both labeled and unlabeled samples for model training, can overcome this problem well. In this paper, a semi-supervised learning method using the convolutional neural network (CNN) is proposed for steel surface defect recognition. The proposed method requires fewer labeled samples, and the unlabeled data can be used to help training. And, the CNN is improved by Pseudo-Label. The experimental results on a benchmark dataset of steel surface defect recognition indicate that the proposed method can achieve good performances with limited labeled data, which achieves an accuracy of 90.7% with 17.53% improvement. Furthermore, the proposed method has been applied to a real-world case from a Chinese steel company, and obtains an accuracy of 86.72% which significantly better than the original method in this workshop.