Semisupervised learning for a hybrid generative/discriminative classifier based on the maximum entropy principle

This paper presents a method for designing semi-supervised classifiers trained on labeled and unlabeled samples. We focus on probabilistic semi-supervised classifier design for multi-class and single-labeled classification problems, and propose a hybrid approach that takes advantage of generative and discriminative approaches. In our approach, we first consider a generative model trained by using labeled samples and introduce a bias correction model, where these models belong to the same model family, but have different parameters. Then, we construct a hybrid classifier by combining these models based on the maximum entropy principle. To enable us to apply our hybrid approach to text classification problems, we employed naive Bayes models as the generative and bias correction models. Our experimental results for four text data sets confirmed that the generalization ability of our hybrid classifier was much improved by using a large number of unlabeled samples for training when there were too few labeled samples to obtain good performance. We also confirmed that our hybrid approach significantly outperformed generative and discriminative approaches when the performance of the generative and discriminative approaches was comparable. Moreover, we examined the performance of our hybrid classifier when the labeled and unlabeled data distributions were different.     

Exploiting local structure in Boltzmann machines

Restricted Boltzmann machines (RBM) are well-studied generative models. For image data, however, standard RBMs are suboptimal, since they do not exploit the local nature of image statistics. We modify RBMs to focus on local structure by restricting visible-hidden interactions. We model long-range dependencies using direct or indirect lateral interaction between hidden variables. While learning in our model is much faster, it retains generative and discriminative properties of RBMs of similar complexity.     

On estimating simple probabilistic discriminative models with subclasses

Discriminative subclass models can provide good estimates of complex ‘continuous to discrete’ conditional probabilities for hybrid Bayesian network models. However, the conventional approach of specifying deterministic ‘hard’ subclasses via unsupervised clustering can lead to inaccurate models. The multimodal softmax (MMS) model is presented as a new probabilistic discriminative subclass model that overcomes this unreliability. By invoking fully probabilistic latent ‘soft’ subclasses, MMS permits learning via standard statistical methods without requiring explicit clustering/relabeling of data. MMS is also shown to be closely related to the mixture of experts model and the generative Gaussian mixture classifier. Synthetic and benchmark classification results demonstrate the MMS model’s correctness and usefulness for hybrid probabilistic modeling.     

A multi-class classification strategy for Fisher scores: Application to signer independent sign language recognition

Fisher kernels combine the powers of discriminative and generative classifiers by mapping the variable-length sequences to a new fixed length feature space, called the Fisher score space. The mapping is based on a single generative model and the classifier is intrinsically binary. We propose a multi-class classification strategy that applies a multi-class classification on each Fisher score space and combines the decisions of multi-class classifiers. We experimentally show that the Fisher scores of one class provide discriminative information for the other classes as well. We compare several multi-class classification strategies for Fisher scores generated from the hidden Markov models of sign sequences. The proposed multi-class classification strategy increases the classification accuracy in comparison with the state of the art strategies based on combining binary classifiers. To reduce the computational complexity of the Fisher score extraction and the training phases, we also propose a score space selection method and show that, similar or even higher accuracies can be obtained by using only a subset of the score spaces. Based on the proposed score space selection method, a signer adaptation technique is also presented that does not require any re-training.     

Boosted Bayesian network classifiers

The use of Bayesian networks for classification problems has received a significant amount of recent attention. Although computationally efficient, the standard maximum likelihood learning method tends to be suboptimal due to the mismatch between its optimization criteria (data likelihood) and the actual goal of classification (label prediction accuracy). Recent approaches to optimizing classification performance during parameter or structure learning show promise, but lack the favorable computational properties of maximum likelihood learning. In this paper we present boosted Bayesian network classifiers, a framework to combine discriminative data-weighting with generative training of intermediate models. We show that boosted Bayesian network classifiers encompass the basic generative models in isolation, but improve their classification performance when the model structure is suboptimal. We also demonstrate that structure learning is beneficial in the construction of boosted Bayesian network classifiers. On a large suite of benchmark data-sets, this approach outperforms generative graphical models such as naive Bayes and TAN in classification accuracy. Boosted Bayesian network classifiers have comparable or better performance in comparison to other discriminatively trained graphical models including ELR and BNC. Furthermore, boosted Bayesian networks require significantly less training time than the ELR and BNC algorithms.     

Discriminative Supervised Hashing for Cross-Modal Similarity Search

With the advantages of low storage cost and high retrieval efficiency, hashing techniques have recently been an emerging topic in cross-modal similarity search. As multiple modal data reflect similar semantic content, many works aim at learning unified binary codes. However, discriminative hashing features learned by these methods are not adequate. This results in lower accuracy and robustness. We propose a novel hashing learning framework which jointly performs classifier learning, subspace learning, and matrix factorization to preserve class-specific semantic content, termed Discriminative Supervised Hashing (DSH), to learn the discriminative unified binary codes for multi-modal data. Besides, reducing the loss of information and preserving the non-linear structure of data, DSH non-linearly projects different modalities into the common space in which the similarity among heterogeneous data points can be measured. Extensive experiments conducted on the three publicly available datasets demonstrate that the framework proposed in this paper outperforms several state-of-the-art methods.     

Adaptive semi-supervised learning on labeled and unlabeled data with different distributions

Developing methods for designing good classifiers from labeled samples whose distribution is different from that of test samples is an important and challenging research issue in the fields of machine learning and its application. This paper focuses on designing semi-supervised classifiers with a high generalization ability by using unlabeled samples drawn by the same distribution as the test samples and presents a semi-supervised learning method based on a hybrid discriminative and generative model. Although JESS-CM is one of the most successful semi-supervised classifier design frameworks based on a hybrid approach, it has an overfitting problem in the task setting that we consider in this paper. We propose an objective function that utilizes both labeled and unlabeled samples for the discriminative training of hybrid classifiers and then expect the objective function to mitigate the overfitting problem. We show the effect of the objective function by theoretical analysis and empirical evaluation. Our experimental results for text classification using four typical benchmark test collections confirmed that with our task setting in most cases, the proposed method outperformed the JESS-CM framework. We also confirmed experimentally that the proposed method was useful for obtaining better performance when classifying data samples into either known or unknown classes, which were included in given labeled samples or not, respectively.     

产生式与判别式线性混合分类器

产生式方法和判别式方法是解决分类问题的两种不同框架,具有各自的优势。为利用两种方法各自的优势,文中提出一种产生式与判别式线性混合分类模型,并设计一种基于遗传算法的产生式与判别式线性混合分类模型的学习算法。该算法将线性混合分类器混合参数的学习看作一个最优化问题,以两个基分类器对每个训练数据的后验概率值为数据依据,用遗传算法找出线性混合分类器混合参数的最优值。实验结果表明,在大多数数据集上,产生式与判别式线性混合分类器的分类准确率优于或近似于它的两个基分类器中的优者。     

多文化场景下的多模态情感识别

自动情感识别是一个非常具有挑战性的课题,并且有着广泛的应用价值.本文探讨了在多文化场景下的多模态情感识别问题.我们从语音声学和面部表情等模态分别提取了不同的情感特征,包括传统的手工定制特征和基于深度学习的特征,并通过多模态融合方法结合不同的模态,比较不同单模态特征和多模态特征融合的情感识别性能.我们在CHEAVD中文多模态情感数据集和AFEW英文多模态情感数据集进行实验,通过跨文化情感识别研究,我们验证了文化因素对于情感识别的重要影响,并提出3种训练策略提高在多文化场景下情感识别的性能,包括：分文化选择模型、多文化联合训练以及基于共同情感空间的多文化联合训练,其中基于共同情感空间的多文化联合训练通过将文化影响与情感特征分离,在语音和多模态情感识别中均取得最好的识别效果.     

混合生成式和判别式模型的图像自动标注

目的 由于图像检索中存在着低层特征和高层语义之间的“语义鸿沟”,图像自动标注成为当前的关键性问题.为缩减语义鸿沟,提出了一种混合生成式和判别式模型的图像自动标注方法.方法 在生成式学习阶段,采用连续的概率潜在语义分析模型对图像进行建模,可得到相应的模型参数和每幅图像的主题分布.将这个主题分布作为每幅图像的中间表示向量,那么图像自动标注的问题就转化为一个基于多标记学习的分类问题.在判别式学习阶段,使用构造集群分类器链的方法对图像的中间表示向量进行学习,在建立分类器链的同时也集成了标注关键词之间的上下文信息,因而能够取得更高的标注精度和更好的检索效果.结果 在两个基准数据集上进行的实验表明,本文方法在Corel5k数据集上的平均精度、平均召回率分别达到0.28和0.32,在IAPR-TC12数据集上则达到0.29和0.18,其性能优于大多数当前先进的图像自动标注方法.此外,从精度—召回率曲线上看,本文方法也优于几种典型的具有代表性的标注方法.结论 提出了一种基于混合学习策略的图像自动标注方法,集成了生成式模型和判别式模型各自的优点,并在图像语义检索的任务中表现出良好的有效性和鲁棒性.本文方法和技术不仅能应用于图像检索和识别的领域,经过适当的改进之后也能在跨媒体检索和数据挖掘领域发挥重要作用.     

Infrared ultraspectral signature classification based on a restricted Boltzmann machine with sparse and prior constraints

The state-of-the-art ultraspectral technology brings a new hope for the high precision applications due to its high spectral resolution. However, it comes with new challenges brought by the improvement of spectral resolution such as the Hughes phenomenon and over-fitting issue, and our work is aimed at addressing these problems. As new Markov random field (MRF) models, the restricted Boltzmann machines (RBMs) have been used as generative models for many different pattern recognition and artificial intelligence applications showing promising and outstanding performance. In this article, we propose a new method for infrared ultraspectral signature classification based on the RBMs, which adopt the regularization-based techniques to improve the classification accuracy and robustness to noise compared to traditional RBMs. First, we add an arctan-like term to the objective function as a sparse constraint to improve the classification accuracy. Second, we utilize a Gaussian prior to avoid the over-fitting problem. Third, to further improve the classification performance, a multi-layer RBM model, a deep belief network (DBN), is adopted for infrared ultraspectral signature classification. Experiments using different spectral libraries provided by the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and the Environmental Protection Agency (EPA) were performed to evaluate the performance of the proposed method by comparing it with other traditional methods, including spectral coding-based classifiers (binary coding (BC), spectral feature-based binary coding (SFBC), and spectral derivative feature coding (SDFC) matching methods), a novel feature extraction method termed crosscut feature extraction matching (CF), and three machine learning methods (artificial deoxyribonucleic acid (DNA)-based spectral matching (ADSM), DBN, and sparse deep belief network (SparseDBN)). Experimental results demonstrate that the proposed method is superior to the other methods with which it was compared and can simultaneously improve the accuracy and robustness of classification.     

基于Wasserstein GAN的文档表示模型

文档表示模型可以将非结构化的文本数据转化为结构化数据,是多种自然语言处理任务的基础,而目前基于词的模型在文档表示任务中有着无法直接表示文档的缺陷。针对此问题,基于生成对抗网络GAN可以使用两个神经网络进行对抗学习,从而很好地学习到原始数据分布的特点,提出了文档表示模型WADM,使用去噪自编码器作为其判别网络,由其隐层直接得到文档的分布表示。实验表明,WADM能够准确抽取文档特征,相比基于词的模型具有更强的文档表示能力。     

Learning semantic concepts from image database with hybrid generative/discriminative approach

Semantic gap has become a bottleneck of content-based image retrieval in recent years. In order to bridge the gap and improve the retrieval performance, automatic image annotation has emerged as a crucial problem. In this paper, a hybrid approach is proposed to learn the semantic concepts of images automatically. Firstly, we present continuous probabilistic latent semantic analysis (PLSA) and derive its corresponding Expectation–Maximization (EM) algorithm. Continuous PLSA assumes that elements are sampled from a multivariate Gaussian distribution given a latent aspect, instead of a multinomial one in traditional PLSA. Furthermore, we propose a hybrid framework which employs continuous PLSA to model visual features of images in generative learning stage and uses ensembles of classifier chains to classify the multi-label data in discriminative learning stage. Therefore, the framework can learn the correlations between features as well as the correlations between words. Since the hybrid approach combines the advantages of generative and discriminative learning, it can predict semantic annotation precisely for unseen images. Finally, we conduct the experiments on three baseline datasets and the results show that our approach outperforms many state-of-the-art approaches.     

Discriminative learning can succeed where generative learning fails

Generative algorithms for learning classifiers use training data to separately estimate a probability model for each class. New items are classified by comparing their probabilities under these models. In contrast, discriminative learning algorithms try to find classifiers that perform well on all the training data.We show that there is a learning problem that can be solved by a discriminative learning algorithm, but not by any generative learning algorithm. This statement is formalized using a framework inspired by previous work of Goldberg [P. Goldberg, When can two unsupervised learners achieve PAC separation?, in: Proceedings of the 14th Annual COLT, 2001, pp. 303-319].     

Hybrid deep neural network model for human action recognition

In this paper, we propose a hybrid deep neural network model for recognizing human actions in videos. A hybrid deep neural network model is designed by the fusion of homogeneous convolutional neural network (CNN) classifiers. The ensemble of classifiers is built by diversifying the input features and varying the initialization of the weights of the neural network. The convolutional neural network classifiers are trained to output a value of one, for the predicted class and a zero, for all the other classes. The outputs of the trained classifiers are considered as confidence value for prediction so that the predicted class will have a confidence value of approximately 1 and the rest of the classes will have a confidence value of approximately 0. The fusion function is computed as the maximum value of the outputs across all classifiers, to pick the correct class label during fusion. The effectiveness of the proposed approach is demonstrated on UCF50 dataset resulting in a high recognition accuracy of 99.68%.     

基于判别式受限玻尔兹曼机的医学图像分类法

随着计算机技术的发展,越来越多的医学图像分析技术应运而生.利用数据挖掘方法对医学图像做分析是目前研究的热点之一,该方法首先从医学图像中提取统计特征,在此基础上进一步挖掘,这种方法对所提取的特征有很强的依赖性而且受到经验等主观因素的影响.针对乳腺X光图像,采用一种可以从图像中自动学习特征并利用学习到的特征对图像进行分类的医学图像分析新方法——判别式受限玻尔兹曼机(Discriminative Restricted Boltzmann Machine,DRBM).DRBM是一种无向判别模型,它可以自动地从图像中学习特征.在乳腺X光图像标准数据集上的实验结果表明,DRBM对医学图像的分类准确率明显高于其它基于统计特征提取的医学图像分类方法.     

Learning the best subset of local features for face recognition

We propose a novel, local feature-based face representation method based on two-stage subset selection where the first stage finds the informative regions and the second stage finds the discriminative features in those locations. The key motivation is to learn the most discriminative regions of a human face and the features in there for person identification, instead of assuming a priori any regions of saliency. We use the subset selection-based formulation and compare three variants of feature selection and genetic algorithms for this purpose. Experiments on frontal face images taken from the FERET dataset confirm the advantage of the proposed approach in terms of high accuracy and significantly reduced dimensionality.     

一种基于属性分割的产生式/判别式混合分类器

为了利用产生式和判别式方法各自的优势,研究了基于属性分割的产生式/判别式混合分类模型框架,提出了一种基于属性分割的产生式/判别式混合分类器学习算法GDGA。其利用遗传算法,将属性集X划分为两个子集XG和XD,并相应地将训练集D垂直分割为两个子集DG和DD,在两个训练子集上分别学习产生式分类器和判别式分类器;最后将两个分类器合并形成一个混合分类器。实验结果表明,在大多数数据集上,混合分类器的分类正确率优于其成员分类器。在训练数据不足或数据属性分布不清楚的情况下,该混合分类器具有特别的优势。     

一种能量函数意义下的生成式对抗网络

生成式对抗网络（Generative adversarial network,GAN）是目前人工智能领域的一个研究热点,引起了众多学者的关注.针对现有GAN生成模型效率低下和判别模型的梯度消失问题,本文提出一种基于重构误差的能量函数意义下的生成式对抗网络模型（Energy reconstruction error GAN,E-REGAN）.首先,将自适应深度信念网络（Adaptive deep belief network,ADBN）作为生成模型,来快速学习给定样本数据的概率分布并进一步生成相似的样本数据.其次,将自适应深度自编码器（Adaptive deep autoencoder,ADAE）的重构误差（Reconstruction error,RE）作为一个表征判别模型性能的能量函数,能量越小表示GAN学习优化过程越趋近纳什均衡的平衡点,否则反之.同时,通过反推法给出了E-REGAN的稳定性分析.最后在MNIST和CIFAR-10标准数据集上的实验结果表明,相较于现有的类似模型,E-REGAN在学习速度和数据生成能力两方面均有较大提高.