Machine learning and deep learning

Electronic Markets - Today, intelligent systems that offer artificial intelligence capabilities often rely on machine learning. Machine learning describes the capacity of systems to learn from...     

结合集成学习与迁移学习的标签比例学习方法

标签比例学习（LLP）是一种将实例放入包中的机器学习方法,它只提供包中的实例信息和标签比例信息,而不提供标签信息。针对多个相关任务的LLP问题,提出了一种基于迁移学习的标签比例集成学习模型,简称AT-LLP,该模型通过在任务之间构建共享参数来连接相关任务,将源任务中学习到的知识迁移到目标任务中,从而提高目标任务的学习效率。同时该算法引入了集成学习算法,在分类器多轮迭代的学习过程中,不断调整训练集的权重系数,进一步将弱分类器训练为强分类器。实验表明,所提AT-LLP模型比现有LLP方法具有更好的性能。     

Machine learning by imitating human learning

Learning general concepts in imperfect environments is difficult since training instances often include noisy data, inconclusive data, incomplete data, unknown attributes, unknown attribute values and other barriers to effective learning. It is well known that people can learn effectively in imperfect environments, and can manage to process very large amounts of data. Imitating human learning behavior therefore provides a useful model for machine learning in real-world applications. This paper proposes a new, more effective way to represent imperfect training instances and rules, and based on the new representation, a Human-Like Learning (HULL) algorithm for incrementally learning concepts well in imperfect training environments. Several examples are given to make the algorithm clearer. Finally, experimental results are presented that show the proposed learning algorithm works well in imperfect learning environments.     

Part 4: Reinforcement learning: Machine learning and natural learning

The theory of reinforcement learning (RL) was originally motivated by animal learning of sequential behavior, but has been developed and extended in the field of machine learning as an approach to Markov decision processes. Recently, a number of neuroscience studies have suggested a relationship between reward-related activities in the brain and functions necessary for RL. Regarding the history of RL, we introduce in this article the theory of RL and present two engineering applications. Then we discuss possible implementations in the brain.     

Semi-supervised learning combining co-training with active learning

Co-training is a good paradigm of semi-supervised, which requires the data set to be described by two views of features. There are a notable characteristic shared by many co-training algorithm: the selected unlabeled instances should be predicted with high confidence, since a high confidence score usually implies that the corresponding prediction is correct. Unfortunately, it is not always able to improve the classification performance with these high confidence unlabeled instances. In this paper, a new semi-supervised learning algorithm was proposed combining the benefits of both co-training and active learning. The algorithm applies co-training to select the most reliable instances according to the two criterions of high confidence and nearest neighbor for boosting the classifier, also exploit the most informative instances with human annotation for improve the classification performance. Experiments on several UCI data sets and natural language processing task, which demonstrate our method achieves more significant improvement for sacrificing the same amount of human effort.     

Training learning strategies with computer-aided cooperative learning

This paper describes the combination of two potent training technologies (computer-based instruction and cooperative learning) into a system called computer-aided cooperative learning (CACL) and the use of CACL to train students in a general learning strategy. This six-steps strategy involves setting a task-appropriate mood, reading for general understanding, recall as much of the material as possible, detecting errors and omissions, elaborating upon the material to make it more memorable and reviewing. CACL capitalizes on the strengths and overcomes some of the weaknesses of each of the constituent technologies. The resulting program is described and some data demonstrating its effectiveness is presented. Students using CACL recalled more material from each of two passages studied individually than did students who did not use CACL. CACL appears to be a promising technology for the delivery of learning strategies. Future research and development efforts should examine CACL's usefulness to the training of more sophisticated learning strategies.     

初态学习下的迭代学习控制

提出一种新的初态学习律,以放宽常规迭代学习控制方法的初始定位条件．它允许一定的定位误差,在迭代中不需要定位在某一具体位置上,使得学习控制系统具有鲁棒收敛性．针对二阶LTI系统,给出了输入学习律及初态学习律的收敛性充分条件．依据收敛性条件,学习增益的选取需系统矩阵的估计值,但在一定建模误差下,仍能保证算法的收敛性．所提出的初态学习律本身及其收敛性条件均与输入矩阵无关．     

Multiagent learning using a variable learning rate

Learning to act in a multiagent environment is a difficult problem since the normal definition of an optimal policy no longer applies. The optimal policy at any moment depends on the policies of the other agents. This creates a situation of learning a moving target. Previous learning algorithms have one of two shortcomings depending on their approach. They either converge to a policy that may not be optimal against the specific opponents' policies, or they may not converge at all. In this article we examine this learning problem in the framework of stochastic games. We look at a number of previous learning algorithms showing how they fail at one of the above criteria. We then contribute a new reinforcement learning technique using a variable learning rate to overcome these shortcomings. Specifically, we introduce the WoLF principle, “Win or Learn Fast”, for varying the learning rate. We examine this technique theoretically, proving convergence in self-play on a restricted class of iterated matrix games. We also present empirical results on a variety of more general stochastic games, in situations of self-play and otherwise, demonstrating the wide applicability of this method.     

极限学习机类不平衡数据学习算法研究

尽管极限学习机因具有快速、简单、易实现及普适的逼近能力等特点被广泛应用于分类、回归及特征学习问题,但是,极限学习机同其他标准分类方法一样将最大化各类总分类性能作为算法的优化目标,因此,在实际应用中遇到数据样本分布不平衡时,算法对大类样本具有性能偏向性。针对极限学习机类不平衡学习问题的研究起步晚,算法少的问题,在介绍了极限学习机类不平衡数据学习研究现状,极限学习机类不平衡数据学习的典型算法-加权极限学习机及其改进算法的基础上,提出一种不需要对原始不平衡样本进行处理的Adaboost提升的加权极限学习机,通过在15个UCI不平衡数据集进行分析实验,实验结果表明提出的算法具有更好的分类性能。     

概念驱动的小样本判别特征学习方法

小样本学习旨在让模型能够在仅有少量标记数据的新类中进行分类。基于度量学习的方法是小样本学习的一种有效方法,该类方法利用有标签的支持集样本构建类表示,再基于查询样本和类表示的相似性进行分类。因此,如何构建判别性更强的类表示是这类方法的关键所在。多数工作在构建类表示时,忽略了类概念相关信息的挖掘,这样容易引入样本中类别无关信息,从而降低类表示的判别性。为此本文提出一种概念驱动的小样本判别特征学习方法。该方法首先利用类别的语义信息来指导模型挖掘样本中类概念相关信息,进而构建更具判别性的类表示。其次,设计了随机掩码混合机制增加样本的多样性和识别难度,进一步提升类表示的质量。最后对处于决策边界附近的查询样本赋予更大的权重,引导模型关注难样本,从而更好地进行类表示学习。大量实验的结果表明本文提出的方法能够有效提升小样本分类任务的准确率,并且在多个数据集上优于当前先进的算法。     

Imitation guided learning in learning classifier systems

In this paper, we study the means of developing an imitation process allowing to improve learning in the framework of learning classifier systems. We present three different approaches in the way a behavior observed may be taken into account through a guidance interaction: two approaches using a model of this behavior, and one without modelling. Those approaches are evaluated and compared in different environments when they are applied to three major classifier systems: ZCS, XCS and ACS. Results are analyzed and discussed. They highlight the importance of using a model of the observed behavior to enable an efficient imitation. Moreover, they show the advantages of taking this model into account by a specialized internal action. Finally, they bring new results of comparison between ZCS, XCS and ACS.

Relations between Gold-style learning and query learning

Different formal learning models address different aspects of human learning. Below we compare Gold-style learning—modelling learning as a limiting process in which the learner may change its mind arbitrarily often before converging to a correct hypothesis—to learning via queries—modelling learning as a one-shot process in which the learner is required to identify the target concept with just one hypothesis. In the Gold-style model considered below, the information presented to the learner consists of positive examples for the target concept, whereas in query learning, the learner may pose a certain kind of queries about the target concept, which will be answered correctly by an oracle (called teacher). Although these two approaches seem rather unrelated at first glance, we provide characterisations of different models of Gold-style learning (learning in the limit, conservative inference, and behaviourally correct learning) in terms of query learning. Thus we describe the circumstances which are necessary to replace limit learners by equally powerful one-shot learners. Our results are valid in the general context of learning indexable classes of recursive languages. This analysis leads to an important observation, namely that there is a natural query learning type hierarchically in-between Gold-style learning in the limit and behaviourally correct learning. Astonishingly, this query learning type can then again be characterised in terms of Gold-style inference.     

Promoting self-regulated learning in web-based learning environments

Self-regulated learning with the Internet or hypermedia requires not only cognitive learning strategies, but also specific and general meta-cognitive strategies. The purposes of the Study2000 project, carried out at the TU Dresden, were to develop and evaluate authoring tools that support teachers and students in web-based learning and instruction. This paper presents how the authoring tools of the Study2000 project can implement psychologically sound measures to promote (a) active and elaborated learning activities and (b) meta-cognitive activities in a web-based learning environment. Furthermore, it describes a study involving 72 university students in the use of such a web-based learning environment in a self-regulated learning setting at the university level. Results show that students spent almost 70% of their study time with texts, 11% with learning tasks and 12% with the active and elaborated learning tools, whereas meta-cognitive aids where hardly used (<1%).     

Advanced machine learning approaches to personalise learning: learning analytics and decision making

The aim of the paper is to present methodology to personalise learning using learning analytics and to make further decisions on suitability, acceptance and use of personalised learning units. In the paper, first of all, related research review is presented. Further, an original methodology to personalise learning applying learning analytics in virtual learning environments and empirical research results are presented. Using this learning personalisation methodology, decision-making model and method are proposed to evaluate suitability, acceptance and use of personalised learning units. Personalised learning units evaluation methodology presented in the paper is based on (1) well-known principles of Multiple Criteria Decision Analysis for identifying evaluation criteria; (2) Educational Technology Acceptance & Satisfaction Model (ETAS-M) based on well-known Unified Theory on Acceptance and Use of Technology (UTAUT) model, and (3) probabilistic suitability indexes to identify learning components’ suitability to particular students’ needs according to their learning styles. In the paper, there are also examples of implementing the methodology using different weights of evaluation criteria. This methodology is applicable in real life situations where teachers have to help students to create and apply learning units that are most suitable for their needs and thus to improve education quality and efficiency.     

Layered learning design: Towards an integration of learning design and learning object perspectives

The use of ICT to enhance teaching and learning depends on effective design, which operates at many levels of granularity from the small to the very large. This reflects the range of educational problems from course design down to the design of activities focused on specific learning objectives. For maximum impact these layers of design need to be co-ordinated effectively. This paper delineates a reference model of ‘layered learning design’ where designs at one layer should use and incorporate designs from lower (more specific) layers in elegant and powerful ways. This would allow different designers, or tutors, to focus on different levels of abstraction in the learning design process, and to collaborate in combining designs to make a substantial impact on practice.     

Developing an appropriate design of blended learning with web-enabled self-regulated learning to enhance students' learning and thoughts regarding online learning

The vocational schools in Taiwan regard professional certifications as a badge of skills achievement. The teaching in this context usually focuses on how to help students enhance their professional skills and pass the certificate examinations, particularly for computing courses. However, due to national education policy, pure online courses are not permitted here and in some other nations. In order to provide an appropriate design and arrangement of blended learning (BL) courses, the authors redesigned a course, integrating web-enabled self-regulated learning (SRL) with variations in online class frequency, and explored their effects on enhancing students' skills of deploying database management system (DBMS) and their thoughts regarding blended course and interventions concerning SRL. Three class sections with a total of 112 students were taken as three distinct groups. The results indicated that students in the group of SRL and BL with five online classes had the highest grades for using DBMS among the three groups, and had very positive thoughts regarding the interventions concerning BL and SRL. The authors also provide suggestions and implications for teachers and schools to adopt innovative teaching methods and technologies, and redesign their courses to help students learn.     

Flexible learning itinerary vs. linear learning itinerary

The latest video game and entertainment technology and other technologies are facilitating the development of new and powerful e-Learning systems. In this paper, we present a computer-based game for learning about five historical ages. The objective of the game is to reinforce the events that mark the transition from one historical age to another and the order of the historical ages. Our game incorporates natural human–computer interaction based on video game technology, Frontal Projection, and personalized learning. For personalized learning, a Flexible Learning Itinerary has been included, where the children can decide how to direct the flow of their own learning process. For comparison, a Linear Learning Itinerary has also been included, where the children follow a determined learning flow. A study to compare the two different learning itineraries was carried out. Twenty nine children from 8 to 9 years old participated in the study. The analysis of the pre-tests and the post-tests determined that children learned the contents of a game about historical ages. The results show that there were no statistically significant differences between the two learning itineraries. Therefore, our study reveals the potential of computer-based learning games as a tool in the learning process for both flexible and linear itineraries.