Domain transfer multiple kernel learning

Cross-domain learning methods have shown promising results by leveraging labeled patterns from the auxiliary domain to learn a robust classifier for the target domain which has only a limited number of labeled samples. To cope with the considerable change between feature distributions of different domains, we propose a new cross-domain kernel learning framework into which many existing kernel methods can be readily incorporated. Our framework, referred to as Domain Transfer Multiple Kernel Learning (DTMKL), simultaneously learns a kernel function and a robust classifier by minimizing both the structural risk functional and the distribution mismatch between the labeled and unlabeled samples from the auxiliary and target domains. Under the DTMKL framework, we also propose two novel methods by using SVM and prelearned classifiers, respectively. Comprehensive experiments on three domain adaptation data sets (i.e., TRECVID, 20 Newsgroups, and email spam data sets) demonstrate that DTMKL-based methods outperform existing cross-domain learning and multiple kernel learning methods.     

Learning to classify e-mail

In this paper we study supervised and semi-supervised classification of e-mails. We consider two tasks: filing e-mails into folders and spam e-mail filtering. Firstly, in a supervised learning setting, we investigate the use of random forest for automatic e-mail filing into folders and spam e-mail filtering. We show that random forest is a good choice for these tasks as it runs fast on large and high dimensional databases, is easy to tune and is highly accurate, outperforming popular algorithms such as decision trees, support vector machines and naïve Bayes. We introduce a new accurate feature selector with linear time complexity. Secondly, we examine the applicability of the semi-supervised co-training paradigm for spam e-mail filtering by employing random forests, support vector machines, decision tree and naïve Bayes as base classifiers. The study shows that a classifier trained on a small set of labelled examples can be successfully boosted using unlabelled examples to accuracy rate of only 5% lower than a classifier trained on all labelled examples. We investigate the performance of co-training with one natural feature split and show that in the domain of spam e-mail filtering it can be as competitive as co-training with two natural feature splits.     

结构化集成学习垃圾邮件过滤

为了解决垃圾邮件过滤算法低计算复杂度与高分类准确率之间的矛盾,在多域学习框架下提出一种结构化集成学习思想,它根据文档结构组合多个基分类器的结果以追求更高分类性能.采用邮件文档的字符串特征生成多个轻量基分类器,并采用字符串-频率索引存储标注数据,使得每次更新和查询的时间开销是常数量级.根据邮件文档的多域结构特性,提出历史域分类器效力线性组合权和当前域文档分类能力线性组合权.综合考虑历史域分类器效力和当前域文档分类能力,还提出一种能够提高整体分类准确率的综合线性组合权.在TREC立即全反馈垃圾邮件过滤任务上的实验结果表明:基于综合线性组合权的结构化集成学习方法能够在较短的时间(47.24min)内完成过滤任务,整体性能1-ROCA达到参加TREC2007评测的最优过滤器性能(0.005 5).     

Online active multi-field learning for efficient email spam filtering

Email spam causes a serious waste of time and resources. This paper addresses the email spam filtering problem and proposes an online active multi-field learning approach, which is based on the following ideas: (1) Email spam filtering is an online application, which suggests an online learning idea; (2) Email document has a multi-field text structure, which suggests a multi-field learning idea; and (3) It is costly to obtain a label for a real-world email spam filter, which suggests an active learning idea. The online learner regards the email spam filtering as an incremental supervised binary streaming text classification. The multi-field learner combines multiple results predicted by field classifiers in a novel compound weight schema, and each field classifier calculates the arithmetical average of multiple conditional probabilities calculated from feature strings according to a data structure of string-frequency index. Comparing the current variance of field classifying results with the historical variance, the active learner evaluates the classifying confidence and takes the more uncertain email as the more informative sample for which to request a label. The experimental results show that the proposed approach can achieve the state-of-the-art performance with greatly reduced label requirements and very low space-time costs. The performance of our online active multi-field learning, the standard (1-ROCA)% measurement, even exceeds the full feedback performance of some advanced individual text classification algorithms.     

A spam filtering multi-objective optimization study covering parsimony maximization and three-way classification

Classifier performance optimization in machine learning can be stated as a multi-objective optimization problem. In this context, recent works have shown the utility of simple evolutionary multi-objective algorithms (NSGA-II, SPEA2) to conveniently optimize the global performance of different anti-spam filters. The present work extends existing contributions in the spam filtering domain by using three novel indicator-based (SMS-EMOA, CH-EMOA) and decomposition-based (MOEA/D) evolutionary multi-objective algorithms. The proposed approaches are used to optimize the performance of a heterogeneous ensemble of classifiers into two different but complementary scenarios: parsimony maximization and e-mail classification under low confidence level. Experimental results using a publicly available standard corpus allowed us to identify interesting conclusions regarding both the utility of rule-based classification filters and the appropriateness of a three-way classification system in the spam filtering domain.     

A semi-supervised domain adaptation assembling approach for image classification

Automatic annotation of images is one of the fundamental problems in computer vision applications. With the increasing amount of freely available images, it is quite possible that the training data used to learn a classifier has different distribution from the data which is used for testing. This results in degradation of the classifier performance and highlights the problem known as domain adaptation. Framework for domain adaptation typically requires a classification model which can utilize several classifiers by combining their results to get the desired accuracy. This work proposes depth-based and iterative depth-based fusion methods which are basically rank-based fusion methods and utilize rank of the predicted labels from different classifiers. Two frameworks are also proposed for domain adaptation. The first framework uses traditional machine learning algorithms, while the other works with metric learning as well as transfer learning algorithm. Motivated from ImageCLEF’s 2014 domain adaptation task, these frameworks with the proposed fusion methods are validated and verified by conducting experiments on the images from five domains having varied distributions. Bing, Caltech, ImageNet, and PASCAL are used as source domains and the target domain is SUN. Twelve object categories are chosen from these domains. The experimental results show the performance improvement not only over the baseline system, but also over the winner of the ImageCLEF’s 2014 domain adaptation challenge.     

Combining domain knowledge and machine learning for robust fall detection

This paper presents a method for combining domain knowledge and machine learning (CDKML) for classifier generation and online adaptation. The method exploits advantages in domain knowledge and machine learning as complementary information sources. Whereas machine learning may discover patterns in interest domains that are too subtle for humans to detect, domain knowledge may contain information on a domain not present in the available domain dataset. CDKML has three steps. First, prior domain knowledge is enriched with relevant patterns obtained by machine learning to create an initial classifier. Second, genetic algorithms refine the classifier. Third, the classifier is adapted online on the basis of user feedback using the Markov decision process. CDKML was applied in fall detection. Tests showed that the classifiers developed by CDKML have better performance than machine‐learning classifiers generated on a training dataset that does not adequately represent all real‐life cases of the learned concept. The accuracy of the initial classifier was 10 percentage points higher than the best machine‐learning classifier and the refinement added 3 percentage points. The online adaptation improved the accuracy of the refined classifier by an additional 15 percentage points.     

域自适应学习研究进展

传统的机器学习假设测试样本和训练样本来自同一概率分布. 但当前很多学习场景下训练样本和测试样本可能来自不同的概率分布. 域自 适应学习能够有效地解决训练样本和测试样本概率分布不一致的学习问题,作为 机器学习新出现的研究领域在近几年受到了广泛的关注. 鉴于域自适应学习技术 的重要性,综述了域自适应学习的研究进展. 首先概述了域自适应学习的基本问 题,并总结了近几年出现的重要的域自适应学习方法. 接着介绍了近几年提出的 较为经典的域自适应学习理论和当下域自适应学习的热门研究方向,包括样例加 权域自适应学习、特征表示域自适应学习、参数和特征分解域自适应学习和多 源域自适应学习. 然后对域自适应学习进行了相关的理论分析,讨论了高效的度 量判据,并给出了相应的误差界. 接着对当前域自适应学习在算法、模型结构和 实际应用这三个方面的研究新进展进行了综述. 最后分别探讨了域自适应学习在 特征变换和假设、训练优化、模型和数据表示、NLP 研究中存在的问题这四个方面 的有待进一步解决的问题.     

Learning Kernels for Unsupervised Domain Adaptation with Applications to Visual Object Recognition

Domain adaptation aims to correct the mismatch in statistical properties between the source domain on which a classifier is trained and the target domain to which the classifier is to be applied. In this paper, we address the challenging scenario of unsupervised domain adaptation, where the target domain does not provide any annotated data to assist in adapting the classifier. Our strategy is to learn robust features which are resilient to the mismatch across domains and then use them to construct classifiers that will perform well on the target domain. To this end, we propose novel kernel learning approaches to infer such features for adaptation. Concretely, we explore two closely related directions. In the first direction, we propose unsupervised learning of a geodesic flow kernel (GFK). The GFK summarizes the inner products in an infinite sequence of feature subspaces that smoothly interpolates between the source and target domains. In the second direction, we propose supervised learning of a kernel that discriminatively combines multiple base GFKs. Those base kernels model the source and the target domains at fine-grained granularities. In particular, each base kernel pivots on a different set of landmarks—the most useful data instances that reveal the similarity between the source and the target domains, thus bridging them to achieve adaptation. Our approaches are computationally convenient, automatically infer important hyper-parameters, and are capable of learning features and classifiers discriminatively without demanding labeled data from the target domain. In extensive empirical studies on standard benchmark recognition datasets, our appraches yield state-of-the-art results compared to a variety of competing methods.     

带有双判别器的对抗性领域适应图像分类算法

生成对抗网络的出现将对抗学习的思想引入了机器学习的不同知识体系,带来了全新的发展。对抗性的领域适应算法利用一个共享特征提取器提取域不变表征,一个判别器进行辨别,双方通过对抗性的迭代更新方式达到最优解。在数据来源上,生成对抗网络和领域适应都有极其类似的2个域。在目标函数上,两者都试图追寻一致性。从理论和逻辑结构出发分析两者的内在相似性,尝试利用已成熟的生成对抗网络体系从更深层次进一步提升领域适应性能。通过类比,提出使用2个判别器解决已有对抗性领域适应算法中存在的“模式崩溃”问题,并使用伪标签进行结构上的完善。最后,在标准领域适应任务上的实验表明了本文算法的可行性和有效性。     

An innovative analyser for multi-classifier e-mail classification based on grey list analysis

In this paper, we propose a new technique of e-mail classification based on the analysis of grey list (GL) from the output of an integrated model, which uses multi-classifier classification ensembles of statistical learning algorithms. The GL is the output of a list of classifiers which are not categorized as true positive (TP) nor true negative (TN) but in an unclear status. Many works have been done to filter spam from legitimate e-mails using classification algorithms and substantial performance has been achieved with some amount of false-positive (FP) tradeoffs. However, in spam filtering applications the FP problem is unacceptable in many situations, therefore it is critical to properly classify e-mails in the GL. Our proposed technique uses an innovative analyser for making decisions about the status of these e-mails. It has been shown that the performance of our proposed technique for e-mail classification is much better than the existing systems, in terms of reducing FP problems and improving accuracy.     

A survey of multi-source domain adaptation

In many machine learning algorithms, a major assumption is that the training and the test samples are in the same feature space and have the same distribution. However, for many real applications this assumption does not hold. In this paper, we survey the problem where the training samples and the test samples are from different distributions. This problem can be referred as domain adaptation. The training samples, always with labels, are obtained from what is called source domains, while the test samples, which usually have no labels or only a few labels, are obtained from what is called target domains. The source domains and the target domains are different but related to some extent; the learners can learn some information from the source domains for the learning of the target domains. We focus on the multi-source domain adaptation problem where there is more than one source domain available together with only one target domain. A key issue is how to select good sources and samples for the adaptation. In this survey, we review some theoretical results and well developed algorithms for the multi-source domain adaptation problem. We also discuss some open problems which can be explored in future work.     

一种协同过滤式零次学习方法

深度学习算法在很多有监督学习任务上达到了令人满意的结果，但其依赖于大量标注样本，并且使用特定类别训练的分类器，只能对这些类别进行分类.零次学习希望计算机像人类一样，能够结合历史经验与知识进行推理，无需使用大量新类别样本训练，便可达到识别新类别的效果.发现了零次学习任务存在“冷启动”以及矩阵稀疏两个特点，这些特点在推荐任务中同样存在.受推荐任务启发，将零次图像分类任务建模为矩阵填充问题，借鉴推荐领域中协同过滤算法，将稀疏的样本标签矩阵视为非稀疏的视觉特征矩阵和类别特征矩阵的内积结果，进而实现对新类别样本的分类预测.此外，构建了基于类间语义关联的语义图结构，使用图神经网络进行已知类别和新类别之间的知识迁移，以较小代价为类别学得准确的语义特征.在3个经典零次学习数据集上分别进行传统零次学习和广义零次学习实验，实验结果表明：提出的协同过滤式零次学习方法能够有效提升分类精度，且训练代价较小.     

Domain structure-based transfer learning for cross-domain word representation

Cross-domain word representation aims to learn high-quality semantic representations in an under-resourced domain by leveraging information in a resourceful domain. However, most existing methods mainly transfer the semantics of common words across domains, ignoring the semantic relations among domain-specific words. In this paper, we propose a domain structure-based transfer learning method to learn cross-domain representations by leveraging the relations among domain-specific words. To accomplish this, we first construct a semantic graph to capture the latent domain structure using domain-specific co-occurrence information. Then, in the domain adaptation process, beyond domain alignment, we employ Laplacian Eigenmaps to ensure the domain structure is consistently distributed in the learned embedding space. As such, the learned cross-domain word representations not only capture shared semantics across domains, but also maintain the latent domain structure. We performed extensive experiments on two tasks, namely sentiment analysis and query expansion. The experiment results show the effectiveness of our method for tasks in under-resourced domains.     

Learning adaptation knowledge to improve case-based reasoning

Case-Based Reasoning systems retrieve and reuse solutions for previously solved problems that have been encountered and remembered as cases. In some domains, particularly where the problem solving is a classification task, the retrieved solution can be reused directly. But for design tasks it is common for the retrieved solution to be regarded as an initial solution that should be refined to reflect the differences between the new and retrieved problems. The acquisition of adaptation knowledge to achieve this refinement can be demanding, despite the fact that the knowledge source of stored cases captures a substantial part of the problem-solving expertise. This paper describes an introspective learning approach where the case knowledge itself provides a source from which training data for the adaptation task can be assembled. Different learning algorithms are explored and the effect of the learned adaptations is demonstrated for a demanding component-based pharmaceutical design task, tablet formulation. The evaluation highlights the incremental nature of adaptation as a further reasoning step after nearest-neighbour retrieval. A new property-based classification to adapt symbolic values is proposed, and an ensemble of these property-based adaptation classifiers has been particularly successful for the most difficult of the symbolic adaptation tasks in tablet formulation.     

Online transfer learning by leveraging multiple source domains

Transfer learning aims to enhance performance in a target domain by exploiting useful information from auxiliary or source domains when the labeled data in the target domain are insufficient or difficult to acquire. In some real-world applications, the data of source domain are provided in advance, but the data of target domain may arrive in a stream fashion. This kind of problem is known as online transfer learning. In practice, there can be several source domains that are related to the target domain. The performance of online transfer learning is highly associated with selected source domains, and simply combining the source domains may lead to unsatisfactory performance. In this paper, we seek to promote classification performance in a target domain by leveraging labeled data from multiple source domains in online setting. To achieve this, we propose a new online transfer learning algorithm that merges and leverages the classifiers of the source and target domain with an ensemble method. The mistake bound of the proposed algorithm is analyzed, and the comprehensive experiments on three real-world data sets illustrate that our algorithm outperforms the compared baseline algorithms.     

Cost-sensitive three-way email spam filtering

Email spam filtering is typically treated as a binary classification problem that can be solved by machine learning algorithms. We argue that a three-way decision approach provides a more meaningful way to users for precautionary handling their incoming emails. Three email folders instead of two are produced in a three-way spam filtering system, a suspected folder is added to allow users make further examinations of suspicious emails, thereby reducing the chances of misclassification. Different from existing ternary email spam filtering systems, we focus on two issues that are less studied, that is, the computation of required thresholds to define the three email categories, and the interpretation of the cost-sensitive characteristics of spam filtering. Instead of supplying the thresholds based on intuitive understandings of the levels of tolerance for errors, we systematically calculate the thresholds based on decision-theoretic rough set model. A loss function is interpreted as the costs of making classification decisions. A decision is made for which the overall cost is minimum. Experimental results show that the new approach reduces the error rate of misclassifying a legitimate email to spam and demonstrates a better performance for the cost-sensitivity aspect.     

New Applications of Ensembles of Classifiers

Combination (ensembles) of classifiers is now a well established research line. It has been observed that the predictive accuracy of a combination of independent classifiers excels that of the single best classifier. While ensembles of classifiers have been mostly employed to achieve higher recognition accuracy, this paper focuses on the use of combinations of individual classifiers for handling several problems from the practice in the machine learning, pattern recognition and data mining domains. In particular, the study presented concentrates on managing the imbalanced training sample problem, scaling up some preprocessing algorithms and filtering the training set. Here, all these situations are examined mainly in connection with the nearest neighbour classifier. Experimental results show the potential of multiple classifier systems when applied to those situations.     

基于逻辑回归的多任务域快速分类学习算法

多任务学习通过寻找并共享不同任务域之间的共性特征来完成学习,利用知识迁移加速不同任务域的学习为每个任务域构建一个分类器。提出了一种基于罗杰斯特回归模型的多任务学习方法MTC-LR（Multi-task Coupled Logistic Regression）。“罗杰斯特回归模型”已经被成功应用于单任务分类器上,该模型被众多实验证明是有效的,正是这种方法给人们带来了启示。从理论上证明了通过构造多任务分类器的“开销函数”和“差异性度量函数”,MTC-LR算法可以提高多任务分类器的各自分类精度。相比传统的基于SVM的多任务学习方法,MTC-LR并不依赖于核方法而是通过共轭梯度下降法寻找各个分类器的最优参数。同时MTC-LR与采用“罗杰斯特回归模型”的快速算法CDdual更容易结合,可扩展至大样本的多任务分类学习。正是基于上述发现,为了充分高效利用大样本的多任务域数据,满足大样本的快速运算,在MTC-LR算法的基础上,结合最新的CDdual（The Dual Coordinate Descent Method）算法,提出了MTC-LR的快速算法MTC-LR-CDdual,并对该算法进行了相关的理论分析。将该算法在人工数据集和真实数据集上进行了验证,实验结果表明该算法有着较高的识别率、快速的识别速度和较好的鲁棒性。     

A comparative study for content-based dynamic spam classification using four machine learning algorithms

The growth of email users has resulted in the dramatic increasing of the spam emails during the past few years. In this paper, four machine learning algorithms, which are Naïve Bayesian (NB), neural network (NN), support vector machine (SVM) and relevance vector machine (RVM), are proposed for spam classification. An empirical evaluation for them on the benchmark spam filtering corpora is presented. The experiments are performed based on different training set size and extracted feature size. Experimental results show that NN classifier is unsuitable for using alone as a spam rejection tool. Generally, the performances of SVM and RVM classifiers are obviously superior to NB classifier. Compared with SVM, RVM is shown to provide the similar classification result with less relevance vectors and much faster testing time. Despite the slower learning procedure, RVM is more suitable than SVM for spam classification in terms of the applications that require low complexity.