Joint multiple dictionary learning for tensor sparse coding

Traditional dictionary learning algorithms are used for finding a sparse representation on high dimensional data by transforming samples into a one-dimensional (1D) vector. This 1D model loses the inherent spatial structure property of data. An alternative solution is to employ Tensor Decomposition for dictionary learning on their original structural form —a tensor— by learning multiple dictionaries along each mode and the corresponding sparse representation in respect to the Kronecker product of these dictionaries. To learn tensor dictionaries along each mode, all the existing methods update each dictionary iteratively in an alternating manner. Because atoms from each mode dictionary jointly make contributions to the sparsity of tensor, existing works ignore atoms correlations between different mode dictionaries by treating each mode dictionary independently. In this paper, we propose a joint multiple dictionary learning method for tensor sparse coding, which explores atom correlations for sparse representation and updates multiple atoms from each mode dictionary simultaneously. In this algorithm, the Frequent-Pattern Tree (FP-tree) mining algorithm is employed to exploit frequent atom patterns in the sparse representation. Inspired by the idea of K-SVD, we develop a new dictionary update method that jointly updates elements in each pattern. Experimental results demonstrate our method outperforms other tensor based dictionary learning algorithms.

Particle competition and cooperation to prevent error propagation from mislabeled data in semi-supervised learning

Semi-supervised learning is applied to classification problems where only a small portion of the data items is labeled. In these cases, the reliability of the labels is a crucial factor, because mislabeled items may propagate wrong labels to a large portion or even the entire data set. This paper aims to address this problem by presenting a graph-based (network-based) semi-supervised learning method, specifically designed to handle data sets with mislabeled samples. The method uses teams of walking particles, with competitive and cooperative behavior, for label propagation in the network constructed from the input data set. The proposed model is nature-inspired and it incorporates some features to make it robust to a considerable amount of mislabeled data items. Computer simulations show the performance of the method in the presence of different percentage of mislabeled data, in networks of different sizes and average node degree. Importantly, these simulations reveals the existence of the critical points of the mislabeled subset size, below which the network is free of wrong label contamination, but above which the mislabeled samples start to propagate their labels to the rest of the network. Moreover, numerical comparisons have been made among the proposed method and other representative graph-based semi-supervised learning methods using both artificial and real-world data sets. Interestingly, the proposed method has increasing better performance than the others as the percentage of mislabeled samples is getting larger. © 2012 IEEE.

Combined active and semi-supervised learning using particle walking temporal dynamicscombined active and semi-supervised learning using particle walking temporal dynamics

Both Semi-Supervised Leaning and Active Learning are techniques used when unlabeled data is abundant, but the process of labeling them is expensive and/or time consuming. In this paper, those two machine learning techniques are combined into a single nature-inspired method. It features particles walking on a network built from the data set, using a unique random-greedy rule to select neighbors to visit. The particles, which have both competitive and cooperative behavior, are created on the network as the result of label queries. They may be created as the algorithm executes and only nodes affected by the new particles have to be updated. Therefore, it saves execution time compared to traditional active learning frameworks, in which the learning algorithm has to be executed several times. The data items to be queried are select based on information extracted from the nodes and particles temporal dynamics. Two different rules for queries are explored in this paper, one of them is based on querying by uncertainty approaches and the other is based on data and labeled nodes distribution. Each of them may perform better than the other according to some data sets peculiarities. Experimental results on some real-world data sets are provided, and the proposed method outperforms the semi-supervised learning method, from which it is derived, in all of them.

Semi-supervised learning guided by the modularity measure in complex networks

Semi-supervised learning techniques have gained increasing attention in the machine learning community, as a result of two main factors: (1) the available data is exponentially increasing; (2) the task of data labeling is cumbersome and expensive, involving human experts in the process. In this paper, we propose a network-based semi-supervised learning method inspired by the modularity greedy algorithm, which was originally applied for unsupervised learning. Changes have been made in the process of modularity maximization in a way to adapt the model to propagate labels throughout the network. Furthermore, a network reduction technique is introduced, as well as an extensive analysis of its impact on the network. Computer simulations are performed for artificial and real-world databases, providing a numerical quantitative basis for the performance of the proposed method.

ASYTRAIN: a new methodology for teaching and learning antennas

This paper presents ASYTRAIN, a new tool to teach and learn antennas, based on the use of a modular building kit and a low cost portable antenna measurement system that lets the students design and build different types of antennas and observe their characteristics while learning the insights of the subjects. This tool has a methodology guide for try-and-test project development and, makes the students be active antenna engineers instead of passive learners. This experimental learning method arises their motivation during the antenna courses.

Graph-based Semi-Supervised Learning in Acoustic Modeling for Automatic Speech Recognition

Thesis (Ph.D.)--University of Washington, 2016-06

The effect of mastery learning on a course in accounting information system (AIS): A comparison within public and private universities in Taiwan

The Accounting Information System (AIS) is an important course in the Department of Accounting (DoAc) of universities in Taiwan. This course is required for seniors not only because it meets the needs of the profession, but also because it provides continual study for the department's students.^ The scores of The National College and University Joint Entrance Examination (NUEE) show that students with high learning ability are admitted to public universities with high scores, while those with low learning ability are admitted only to private universities. The same situation has been found by the researcher while teaching an AIS course in DoAc of The Public Chun Shin University (CSU) and The Private Chinese Culture University (CCU).^ The purpose of this study was to determine whether low ability students enrolled in private universities in Taiwan in a mastery learning program could attain the same level as high ability students from public universities enrolled in a traditional program. An experimental design was used. The mastery learning method was used to teach three groups of seniors with low learning ability studying in the DoAc at CCU. The traditional method was used to teach the control group which consisted of senior students of DoAc of CSU with high learning ability. As a part of the mastery learning strategy, a formative test, quizzes, and homework were completed by the experimental group only, while the mid-term examination was completed by both groups as part of the course. The dependent variable was the summative test, the final examination. It was completed by both groups upon the course's completion.^ As predicted, there were significant differences between the two groups' results on the pretest. There were no significant differences between the two groups' results on the posttest. These findings support the hypothesis of the study and reveal the effectiveness of mastery learning strategies with low learning ability students. ^

An incremental multilinear system for human face learning and recognition

This dissertation establishes a novel system for human face learning and recognition based on incremental multilinear Principal Component Analysis (PCA). Most of the existing face recognition systems need training data during the learning process. The system as proposed in this dissertation utilizes an unsupervised or weakly supervised learning approach, in which the learning phase requires a minimal amount of training data. It also overcomes the inability of traditional systems to adapt to the testing phase as the decision process for the newly acquired images continues to rely on that same old training data set. Consequently when a new training set is to be used, the traditional approach will require that the entire eigensystem will have to be generated again. However, as a means to speed up this computational process, the proposed method uses the eigensystem generated from the old training set together with the new images to generate more effectively the new eigensystem in a so-called incremental learning process. In the empirical evaluation phase, there are two key factors that are essential in evaluating the performance of the proposed method: (1) recognition accuracy and (2) computational complexity. In order to establish the most suitable algorithm for this research, a comparative analysis of the best performing methods has been carried out first. The results of the comparative analysis advocated for the initial utilization of the multilinear PCA in our research. As for the consideration of the issue of computational complexity for the subspace update procedure, a novel incremental algorithm, which combines the traditional sequential Karhunen-Loeve (SKL) algorithm with the newly developed incremental modified fast PCA algorithm, was established. In order to utilize the multilinear PCA in the incremental process, a new unfolding method was developed to affix the newly added data at the end of the previous data. The results of the incremental process based on these two methods were obtained to bear out these new theoretical improvements. Some object tracking results using video images are also provided as another challenging task to prove the soundness of this incremental multilinear learning method.

The effects of jigsaw cooperative learning on motivation to learn English at Chung -Hwa Institute of Technology, Taiwan

The Ministry of Education (MOE) of the Republic of China in Taiwan considers English to be one of the keys to raising Taiwan's international competitiveness and requires students attending institutes of technology to receive instruction in English as a foreign language (EFL). This study focused on impacts of the use of cooperative learning as a teaching method on EFL learners. The setting was the English classes of Chung-Hwa Institute of Technology (CHIT). The subjects were 77 students from two classes, majoring in Business Administration. ^ The purpose of this study was to determine the differential effects (i.e., achievement in learning English, motivation orientation and intensity, and attitude concerning English language and culture) on students between the traditional Chinese teaching method and the Jigsaw cooperative learning method at CHIT. ^ The research design for the study was quasi-experimental and descriptive. This study utilized three survey instruments and final exam grades to investigate the effect of Jigsaw on the EFL students' competency in English, and on their attitudes about, and level of motivation toward learning English. The independent variable was the instructional method: one class utilized the Jigsaw approach to cooperative learning while the other utilized the traditional Chinese approach. The dependent variables were academic performance, motivation orientation toward English, motivation intensity toward learning English, and attitude toward learning of English and English culture as determined by final exam and questionnaire scores. The questionnaires and exam were administrated at the beginning and end of the semester. ^ Data analysis indicated that students learning cooperatively had higher final course grades and made more integrative statements on the measure of orientation toward learning English than students who learned using the traditional Chinese methods. Participants who learned using cooperative strategies had more positive attitudes about learning English connected with their desire to associate with English speakers and had more positive attitudes about the learning mechanism they experienced than those instructed though traditional Chinese learning strategies. There were no differences between the groups on the measure of motivation intensity. Recommendations were made to improve the use of the Jigsaw method of cooperative learning through both pedagogical and policy modifications. ^

Inclusion as a Moral Challenge: the Potential of the Konstanz Method of Dilemma Discussion® (KMDD®)

This article presents the Konstanz Method of Dilemma Discussion ® (KMDD ®) and explains the integration of the KMDD ® in ethics lessons. In this paper, some special learning effects of this inclusive teaching and learning method are shown. Furthermore, it investigates the questions of how to achieve more knowledge in ethics lessons by dialogue and how to realize better moral development, particularly by handling of differentiation. Moral education of all participants who are involved in the learning process (learners and teacher alike) is a crucial task of every true inclusion. True inclusion means building optimal learning conditions in keeping with the free will of all participants. Because our society is transforming constantly in both global and demographic aspects, coping with these challenges is mandatory.

Análise do recurso a novas tecnologias no ensino de autistas

A massificação das tecnologias, abrangendo cada vez mais instituições e população em geral, originou uma aproximação da informação e do conhecimento às pessoas, permitindo um novo olhar para esta área, para as novas tecnologias e para o seu uso. Tendo como ponto de partida, a sociedade de informação e o recurso às tecnologias, recursos esses tão patentes actualmente, é possível pensar em criar as condições necessárias para que seja viável evoluir sempre mais, através da imaginação e de ideias criativas. Com base na aproximação das tecnologias que são cada vez mais acessíveis, o presente estudo pretende avaliar se as novas tecnologias, nas suas mais variadas formas, podem ser uma mais valia no ensino de autistas. O autismo começa a ser encarado como um campo a desenvolver estudos e a explorar os conceitos tecnológicos. O estudo aqui apresentado, não pretende ser apenas mais um, mas antes mostrar o trabalho efectuado até à data e o que pode ser realizado, expondo propostas. No sentido de conhecer um pouco melhor os conceitos relacionados com a área do autismo, foi necessário recorrer a pesquisas na Internet, à leitura de livros e a reuniões com pessoas ligadas a essa área, tornando possível uma aquisição de saberes e uma nova visão sobre a temática. Explorados os conceitos iniciais do autismo tornou-se imprescindível a procura dos recursos educativos utilizados nesta área, não só recursos empregados no método tradicional de ensino, como também, os modelos que procuram usar as novas tecnologias. A análise dos recursos existentes permitiu conhecer o panorama actual dos métodos e modelos em uso que conjugado com o levantamento de necessidades efectuado no decorrer deste trabalho, contribuíram para compreender as carências existentes e assim com base nesse conhecimento adquirido efectuar uma proposta de modelo educativo baseado em tecnologias para apoio educativo de autistas.

Aplicação do método de aprendizagem baseado na análise de problemas ao ensino da Fisiopatologia

Educação Médica 1991; 2 (2): 29-40

Decision Support Concerning Demand Response Programs Design and Use – A Conceptual Framework and Simulation Tool

Demand response can play a very relevant role in the context of power systems with an intensive use of distributed energy resources, from which renewable intermittent sources are a significant part. More active consumers participation can help improving the system reliability and decrease or defer the required investments. Demand response adequate use and management is even more important in competitive electricity markets. However, experience shows difficulties to make demand response be adequately used in this context, showing the need of research work in this area. The most important difficulties seem to be caused by inadequate business models and by inadequate demand response programs management. This paper contributes to developing methodologies and a computational infrastructure able to provide the involved players with adequate decision support on demand response programs and contracts design and use. The presented work uses DemSi, a demand response simulator that has been developed by the authors to simulate demand response actions and programs, which includes realistic power system simulation. It includes an optimization module for the application of demand response programs and contracts using deterministic and metaheuristic approaches. The proposed methodology is an important improvement in the simulator while providing adequate tools for demand response programs adoption by the involved players. A machine learning method based on clustering and classification techniques, resulting in a rule base concerning DR programs and contracts use, is also used. A case study concerning the use of demand response in an incident situation is presented.

Unsupervised automatic music genre classification

Trabalho apresentado no âmbito do Mestrado em Engenharia Informática, como requisito parcial para obtenção do grau de Mestre em Engenharia Informática

Digital storytelling: student ID in higher education

Identity achievement is related to personality, as well as cognitive and interpersonal development. In tandem with the deep structural changes that have taken place in society, education must also shift towards a teaching approach focused on learning and the overall development of the student. The integration of technology may be the drive to foster the needed changes. We draw on the literature of multiple subject areas as basis for our work, namely: identity construction and self-representation, within a psychological and social standpoint; Higher Education (HE) in Portugal after Bologna, college student development and other intrinsic relationships, namely the role of emotions and interpersonal relationships in the learning process; the technological evolution of storytelling towards Digital Storytelling (DS) – the Californian model – and its connections to identity and education. Ultimately we propose DS as the aggregator capable of humanizing HE while developing essential skills and competences. Grounded on an interpretative/constructivist paradigm, we implemented a qualitative case study to explore DS in HE. In three attempts to collect student data, we gathered detailed observation notes from two Story Circles; twelve student written reflections; fourteen Digital Stories and detailed observation notes from one Story Show. We carried out three focus groups with teachers where we discussed their perceptions of each student prior to and after watching the Digital Stories, in addition to their opinion on DS in HE as a teaching and learning method and its influence on interpersonal relationships. We sought understandings of the integration of DS to analyze student selfperception and self-representation in HE contexts and intersected our findings with teachers’ perceptions of their students. We compared teachers’ and students’ perspectives, through the analysis of data collected throughout the DS process – Story Circle, Story Creation and Story Show – and triangulated that information with the students’ personal reflections and teacher perceptions. Finally we questioned if and how DS may influence teachers’ perceptions of students. We found participants to be the ultimate gatekeepers in our study. Very few students and teachers voluntarily came forth to take part in the study, confirming the challenge remains in getting participants to see the value and understand the academic rigor of DS. Despite this reluctance, DS proved to be an asset for teachers and students directly and indirectly involved in the study. DS challenges HE contexts, namely teacher established perception of students; student’s own expectations regarding learning in HE; the emotional realm, the private vs. public dichotomy and the shift in educational roles.