Online Learning versus Offline Learning

We present an off-line variant of the mistake-bound model of learning. This is an intermediate model between the on-line learning model (Littlestone, 1988, Littlestone, 1989) and the self-directed learning model (Goldman, Rivest & Schapire, 1993, Goldman & Sloan, 1994). Just like in the other two models, a learner in the off-line model has to learn an unknown concept from a sequence of elements of the instance space on which it makes guess and test trials. In all models, the aim of the learner is to make as few mistakes as possible. The difference between the models is that, while in the on-line model only the set of possible elements is known, in the off-line model the sequence of elements (i.e., the identity of the elements as well as the order in which they are to be presented) is known to the learner in advance. On the other hand, the learner is weaker than the self-directed learner, which is allowed to choose adaptively the sequence of elements presented to him.We study some of the fundamental properties of the off-line model. In particular, we compare the number of mistakes made by the off-line learner on certain concept classes to those made by the on-line and self-directed learners. We give bounds on the possible gaps between the various models and show examples that prove that our bounds are tight.Another contribution of this paper is the extension of the combinatorial tool of labeled trees to a unified approach that captures the various mistake bound measures of all the models discussed. We believe that this tool will prove to be useful for further study of models of incremental learning.     

结合记忆与迁移学习的小样本学习

小样本学习是视觉识别中的一个受关注的领域,旨在通过少量的数据来学习新的视觉概念。为了解决小样本问题,一些元学习方法提出从大量辅助任务中学习可迁移的知识并将其应用于目标任务上。为了更好地对知识进行迁移,提出了一种基于记忆的迁移学习方法。提出一种权重分解策略,将部分权重分解为冻结权重与可学习权重,在迁移学习中通过固定冻结权重,仅更新可学习权重的方式来减少模型需要学习的参数。通过一个额外的记忆模块来存储之前任务的经验,在学习新任务时,这些经验被用来初始化模型的参数状态,以此更好地进行迁移学习。通过在miniImageNet、tieredImageNet以及CUB数据集上的实验结果表明,相对于其他先进的方法,该方法在小样本分类任务上取得了具有竞争力甚至是更好的表现。     

自适应主动半监督学习方法

主动学习从大量无标记样本中挑选样本交给专家标记.现有的批抽样主动学习算法主要受3个限制：（1）一些主动学习方法基于单选择准则或对数据、模型设定假设,这类方法很难找到既有不确定性又有代表性的未标记样本;（2）现有批抽样主动学习方法的性能很大程度上依赖于样本之间相似性度量的准确性,例如预定义函数或差异性衡量;（3）噪声标签问题一直影响批抽样主动学习算法的性能.提出一种基于深度学习批抽样的主动学习方法.通过深度神经网络生成标记和未标记样本的学习表示和采用标签循环模式,使得标记样本与未标记样本建立联系,再回到相同标签的标记样本.这样同时考虑了样本的不确定性和代表性,并且算法对噪声标签具有鲁棒性.在提出的批抽样主动学习方法中,算法使用的子模块函数确保选择的样本集合具有多样性.此外,自适应参数的优化,使得主动学习算法可以自动平衡样本的不确定性和代表性.将提出的主动学习方法应用到半监督分类和半监督聚类中,实验结果表明,所提出的主动学习方法的性能优于现有的一些先进的方法.     

融入Universum学习的度量学习算法

度量学习通过更真实的刻画样本之间的距离,来提高分类和聚类的精度。GMML（Geometric Mean Metric Learning）在学习度量矩阵[A]时,使得在该度量下同类点之间的距离尽可能小,不同类点之间的距离尽可能大。GMML用来学习的训练样本均为目标类数据,而对于现实存在的为数众多的同领域非目标类数据,即Universum数据并未加以利用,不免造成信息的浪费,针对此,提出一种新的度量学习算法--融入Universum学习的GMML（U-GMML）。U-GMML期望得到一个新的度量矩阵[A],使得同类点之间的距离尽可能小,不同类点之间的距离尽可能大,且Universum数据与目标类数据的距离尽可能大,从而使得所学习的度量矩阵[A]更有利于分类。真实数据集上的实验结果验证了该算法的有效性。     

Combining Committee-Based Semi-Supervised Learning and Active Learning

Many data mining applications have a large amount of data but labeling data is usually difficult, expensive, or time consuming, as it requires human experts for annotation. Semi-supervised learning addresses this problem by using unlabeled data together with labeled data in the training process. Co-Training is a popular semi-supervised learning algorithm that has the assumptions that each example is represented by multiple sets of features (views) and these views are sufficient for learning and independent given the class. However, these assumptions are strong and are not satisfied in many real-world domains. In this paper, a single-view variant of Co-Training, called Co-Training by Committee (CoBC) is proposed, in which an ensemble of diverse classifiers is used instead of redundant and independent views. We introduce a new labeling confidence measure for unlabeled examples based on estimating the local accuracy of the committee members on its neighborhood. Then we introduce two new learning algorithms, QBC-then-CoBC and QBC-with-CoBC, which combine the merits of committee-based semi-supervised learning and active learning. The random subspace method is applied on both C4.5 decision trees and 1-nearest neighbor classifiers to construct the diverse ensembles used for semi-supervised learning and active learning. Experiments show that these two combinations can outperform other non committee-based ones.     

基于偏向信息学习的双层强化学习算法

传统的强化学习存在收敛速度慢等问题,结合先验知识预置某些偏向可以加快学习速度.但是当先验知识不正确时又可能导致学习过程不收敛.对此,提出基于偏向信息学习的双层强化学习模型.该模型将强化学习过程和偏向信息学习过程结合起来:偏向信息指导强化学习的行为选择策略,同时强化学习指导偏向信息学习过程.该方法在有效利用先验知识的同时能够消除不正确先验知识的影响.针对迷宫问题的实验表明,该方法能够稳定收敛到最优策略;并且能够有效利用先验知识提高学习效率,加快学习过程的收敛.     

Normalized Learning Rule for Iterative Learning Control

International Journal of Control, Automation and Systems - The iterative learning control (ILC) is attractive for its simple structure, easy implementation. So the ILC is applied to various fields....     

Learning to Rank with Extreme Learning Machine

Relevance ranking has been a popular and interesting topic over the years, which has a large variety of applications. A number of machine learning techniques were successfully applied as the learning algorithms for relevance ranking, including neural network, regularized least square, support vector machine and so on. From machine learning point of view, extreme learning machine actually provides a unified framework where the aforementioned algorithms can be considered as special cases. In this paper, pointwise ELM and pairwise ELM are proposed to learn relevance ranking problems for the first time. In particular, ELM type of linear random node is newly proposed together with kernel version of ELM to be linear as well. The famous publicly available dataset collection LETOR is tested to compare ELM-based ranking algorithms with state-of-art linear ranking algorithms.     

面向小样本学习的嵌入学习方法研究综述

为了解决机器学习在样本量较少的情况下所面临的巨大挑战,研究人员提出了小样本学习的概念.在现有的小样本学习研究工作中,嵌入学习方法取得了不错的效果,引发了大量关注.根据训练特征嵌入函数时结合任务特征信息的方式,将嵌入学习方法划分为单一嵌入模型和混合嵌入模型两大类.依据划分的类别,对现有的嵌入学习方法的研究工作展开进行研究...     

基于学习算子的自学习遗传算法设计

遗传算子是影响遗传算法优化效果的重要因素。针对目前遗传算法研究中忽视个体能动性,没有充分利用进化经验信息的不足,提出反映个体学习能力的学习算子。给出了以个体适应度的变化方向和速度为依据的学习算子设计方法及其计算过程。在此基础上与现有的改进遗传算子结合,提出一种新的改进遗传算法-自学习遗传算法,分析了自学习遗传算法与自适应遗传算法之间在原理上的区别。以一个弹道导弹射程优化问题为算例对算法进行了性能测试,结果表明,在采用相同的改进遗传算子的条件下,学习算子能够以较低的代价提高遗传算法的收敛速度,并获得更好的最终优化结果。     

深度学习中的无监督学习方法综述

从2006年开始,深度神经网络在图像/语音识别、自动驾驶等大数据处理和人工智能领域中都取得了巨大成功,其中无监督学习方法作为深度神经网络中的预训练方法为深度神经网络的成功起到了非常重要的作用. 为此,对深度学习中的无监督学习方法进行了介绍和分析,主要总结了两类常用的无监督学习方法,即确定型的自编码方法和基于概率型受限玻尔兹曼机的对比散度等学习方法,并介绍了这两类方法在深度学习系统中的应用,最后对无监督学习面临的问题和挑战进行了总结和展望.     

基于并行学习的多层极限学习机

为了进一步提高极限学习机的学习性能,将并行学习的思想引入单层极限学习机,并提出了基于并行学习的多层极限学习机模型。实验结果表明,该模型比传统的单层极限学习机、多层极限学习机以及传统基于误差反向学习的深度学习模型分类准确率高、收敛速度快。     

基于主动学习和半监督学习的多类图像分类

多数图像分类算法需要大量的训练样本对分类器模型进行训练.在实际应用中, 对大量样本进行标注非常枯燥、耗时.对于一些特殊图像,如合成孔径雷达 (Synthetic aperture radar, SAR)图像, 对其内容判读非常困难,因此能够获得的标注样本数量非常有限. 本文将基于最优标号和次优标号(Best vs second-best, BvSB)的主动学习和带约束条件的自学习(Constrained self-training, CST) 引入到基于支持向量机(Support vector machine, SVM)分类器的图像分类算法中,提出了一种新的图像分类方法.通过BvSB 主动学习去挖掘那些对当前分类器模型最有价值的样本进行人工标注,并借助CST半 监督学习进一步利用样本集中大量的未标注样本,使得在花费较小标注代价情况下, 能够获得良好的分类性能.将新方法与随机样本选择、基于熵的不确定性采样主动学 习算法以及BvSB主动学习方法进行了性能比较.对3个光学图像集及1个SAR图像集分类 问题的实验结果显示,新方法能够有效地减少分类器训练时所需的人工标注样本的数 量,并获得较高的准确率和较好的鲁棒性.     

一种新的归纳学习算法——基于特征可分性的归纳学习算法

本文提出了一种新的基于特征可分性的归纳学习算法(SBI)。与现有各种归纳学习算法相比,该方法直接从特征对不同类型的可分性出发,建立可分性判据,然后形成决策树,可对多种概念进行判别。SBI算法具有直观且计算简便等优点。本文以实例表明了SBI算法的有效性。     

Learning cultures

For a variety of reasons, learning should be studied as a cultural phenomenon. The task of the first part of this article is to clear up the terminological questions about various ideal types of learning cultures, and how ideal type analysis may be used to study value and knowledge transfer and knowledge acquirement in various types of organisations. The important task of the second part is to analyse how implementation of environmental management systems, like BS-7750, contribute to a certain learning practice. Moreover the task is to evaluate and discuss supplementing or alternative modes of learning to just formalisation of knowledge transfer.     

基于虚拟学习社区的智能学习算法研究

随着信息技术和互联网技术的发展,虚拟学习社区为学习者提供一个开放的学习环境,是满足时代教育要求的强有力的平台,成为学习者网络交流、网络学习的重要手段。如何利用虚拟学习社区来提高学习者的学习效率成为当前研究的热点和重点。本文基于虚拟学习社区,提出一种智能的个性化推荐学习算法,使学习者能够方便、有效、智能、反思性学习,为构建虚拟学习社区提供理论技术支撑,以便更好地指导虚拟学习社区的学习实践。     

一种具有增量学习能力的PU主动学习算法

在正例和无标记样本增量学习中,初始正例样本较少且不同类别正例的反例获取困难,使分类器的分类和泛化能力不强,为解决上述问题,提出一种具有增量学习能力的PU主动学习算法,在使用3个支持向量机进行协同半监督学习的同时,利用基于网格的聚类方法进行无监督学习,当分类与聚类结果不一致时,引入主动学习对无标记样本进行标记。实验结果表明,将该算法应用于Deep Web入口的在线判断和分类能有效提高入口判断的准确性及分类的正确性。     

案例教学与适应性学习之初探

以适应性学习为基础,提出了在自适应性学习(Adaptive Learning)中如何将案例教学(Case-based Learning)融入其中,在学习中如何对学习进行评价,并根据评价为学习者提供与其相适应的案例,让学习者在情境中学习,使得学习更加高效可行。     

Projection Learning

A method of combining learning algorithms is described that preserves attribute-efficiency. It yields learning algorithms that require a number of examples that is polynomial in the number of relevant variables and logarithmic in the number of irrelevant ones. The algorithms are simple to implement and realizable on networks with a number of nodes linear in the total number of variables. They include generalizations of Littlestone's Winnow algorithm, and are, therefore, good candidates for experimentation on domains having very large numbers of attributes but where nonlinear hypotheses are sought.