Active Learning arises as an important issue in several supervised learning scenarios where obtaining data is cheap, but labeling is costly. In general, this consists in a query strategy, a greedy heuristic based on some selection criterion, which searches for the potentially most informative observations to be labeled in order to form a training set. A query strategy is therefore a biased sampling procedure since it systematically favors some observations by generating biased training sets, instead of making independent and identically distributed draws. The main hypothesis of this thesis lies in the reduction of the bias inherited from the selection criterion. The general proposal consists in reducing the bias by selecting the minimal training set from which the estimated probability distribution is as close as possible to the underlying distribution of overall observations. For that, a novel general active learning query strategy has been developed using an Information-Theoretic framework. Several experiments have been performed in order to evaluate the performance of the proposed strategy. The obtained results confirm the hypothesis about the bias, showing that the proposal outperforms the baselines in different datasets.

The field of Machine Learning is concerned with the development of algorithms, models and techniques that solve challenging computational problems by learning from data representative of the problem (e.g. given a set of medical images previously classified by a human expert, build a model to predict unseen images as either benign or malignant). Many important real-world problems have been formulated as supervised learning problems. The assumption is that a data set is available containing the correct output (e.g. class label or target value) for each given data point. In many application domains, obtaining the correct outputs (labels) for data points is a costly and time-consuming task. This has provided the motivation for the development of Machine Learning techniques that attempt to minimize the number of labeled data points while maintaining good generalization performance on a given problem. Active Learning is one such class of techniques and is the focus of this thesis. Active Learning algorithms select or generate unlabeled data points to be labeled and use these points for learning. If successful, an Active Learning algorithm should be able to produce learning performance (e.g test set error) comparable to an equivalent supervised learner using fewer labeled data points. Theoretical, algorithmic and experimental Active Learning research has been conducted and a number of successful applications have been demonstrated. However, the scope of many of the experimental studies on Active Learning has been relatively small and there are very few large-scale experimental evaluations of Active Learning techniques. A significant amount of performance variability exists across Active Learning experimental results in the literature. Furthermore, the implementation details and effects of experimental factors have not been closely examined in empirical Active Learning research, creating some doubt over the strength and generality of conclusions that can be drawn from such results. The Active Learning model/system used in this thesis is the Optimal Active Learning algorithm framework with Gaussian Processes for regression problems (however, most of the research questions are of general interest in many other Active Learning scenarios). Experimental and implementation details of the Active Learning system used are described in detail, using a number of regression problems and datasets of different types. It is shown that the experimental results of the system are subject to significant variability across problem datasets. The hypothesis that experimental factors can account for this variability is then investigated. The results show the impact of sampling and sizes of the datasets used when generating experimental results. Furthermore, preliminary experimental results expose performance variability across various real-world regression problems. The results suggest that these experimental factors can (to a large extent) account for the variability observed in experimental results. A novel resampling technique for Optimal Active Learning, called '3-Sets Cross-Validation', is proposed as a practical solution to reduce experimental performance variability. Further results confirm the usefulness of the technique. The thesis then proposes an extension to the Optimal Active Learning framework, to perform learning via membership queries via a novel algorithm named MQOAL. The MQOAL algorithm employs the Metropolis-Hastings Markov chain Monte Carlo (MCMC) method to sample data points for query selection. Experimental results show that MQOAL provides comparable performance to the pool-based OAL learner, using a very generic, simple MCMC technique, and is robust to experimental factors related to the MCMC implementation. The possibility of making queries in batches is also explored experimentally, with results showing that while some performance degradation does occur, it is minimal for learning in small batch sizes, which is likely to be valuable in some real-world problem domains.

The field of Machine Learning is concerned with the development of algorithms, models and techniques that solve challenging computational problems by learning from data representative of the problem (e.g. given a set of medical images previously classified by a human expert, build a model to predict unseen images as either benign or malignant). Many important real-world problems have been formulated as supervised learning problems. The assumption is that a data set is available containing the correct output (e.g. class label or target value) for each given data point. In many application domains, obtaining the correct outputs (labels) for data points is a costly and time-consuming task. This has provided the motivation for the development of Machine Learning techniques that attempt to minimize the number of labeled data points while maintaining good generalization performance on a given problem. Active Learning is one such class of techniques and is the focus of this thesis. Active Learning algorithms select or generate unlabeled data points to be labeled and use these points for learning. If successful, an Active Learning algorithm should be able to produce learning performance (e.g test set error) comparable to an equivalent supervised learner using fewer labeled data points. Theoretical, algorithmic and experimental Active Learning research has been conducted and a number of successful applications have been demonstrated. However, the scope of many of the experimental studies on Active Learning has been relatively small and there are very few large-scale experimental evaluations of Active Learning techniques. A significant amount of performance variability exists across Active Learning experimental results in the literature. Furthermore, the implementation details and effects of experimental factors have not been closely examined in empirical Active Learning research, creating some doubt over the strength and generality of conclusions that can be drawn from such results. The Active Learning model/system used in this thesis is the Optimal Active Learning algorithm framework with Gaussian Processes for regression problems (however, most of the research questions are of general interest in many other Active Learning scenarios). Experimental and implementation details of the Active Learning system used are described in detail, using a number of regression problems and datasets of different types. It is shown that the experimental results of the system are subject to significant variability across problem datasets. The hypothesis that experimental factors can account for this variability is then investigated. The results show the impact of sampling and sizes of the datasets used when generating experimental results. Furthermore, preliminary experimental results expose performance variability across various real-world regression problems. The results suggest that these experimental factors can (to a large extent) account for the variability observed in experimental results. A novel resampling technique for Optimal Active Learning, called '3-Sets Cross-Validation', is proposed as a practical solution to reduce experimental performance variability. Further results confirm the usefulness of the technique. The thesis then proposes an extension to the Optimal Active Learning framework, to perform learning via membership queries via a novel algorithm named MQOAL. The MQOAL algorithm employs the Metropolis-Hastings Markov chain Monte Carlo (MCMC) method to sample data points for query selection. Experimental results show that MQOAL provides comparable performance to the pool-based OAL learner, using a very generic, simple MCMC technique, and is robust to experimental factors related to the MCMC implementation. The possibility of making queries in batches is also explored experimentally, with results showing that while some performance degradation does occur, it is minimal for learning in small batch sizes, which is likely to be valuable in some real-world problem domains.

This literature review analyses what advantages and disadvantages can be gained by utilizing the flipped classroom method. The Swedish curriculum has been taken into consideration when conducting this review in order to see whether the method is compatible with the aim of the subject of English in the Swedish upper secondary school. Two key principles behind the flipped classroom are ‘active learning’ and ‘student-centered learning’. These key principles have been discussed and contrasted with the terms: ‘passive learning’ and ‘teacher-centered learning’. Acknowledging the fact that it is a challenge to engage students in their own learning as well as have them realize the benefits of doing so, this paper also highlights the advantages of implementing technology tools in the classroom where one can meet the needs of the students of today. The results of this literature review show that there are many advantages when utilizing the flipped classroom method. For example, the flipped classroom creates well-suited environments where discussions can take place, as the in-class activities aim at all students being engaged, applying concepts and sharing ideas. However, the results also show some disadvantages that may come when converting to a flipped classroom. Some of these are that the flipped model takes adjustment time, students need to adapt to the transition to a flipped classroom,and preparing online lessons is time consuming. Furthermore, both students and teachers need to learn how to use the new technology that comes with the flipped classroom model.

abstract: From the instructional perspective, the scope of "active learning" in the literature is very broad and includes all sorts of classroom activities that engage students with the learning experience. However, classifying all classroom activities as a mode of "active learning" simply ignores the unique cognitive processes associated with the type of activity. The lack of an extensive framework and taxonomy regarding the relative effectiveness of these "active" activities makes it difficult to compare and contrast the value of conditions in different studies in terms of student learning. Recently, Chi (2009) proposed a framework of differentiated overt learning activities (DOLA) as active, constructive, and interactive based on their underlying cognitive principles and their effectiveness on students' learning outcomes. The motivating question behind this framework is whether some types of engagement affect learning outcomes more than the others. This work evaluated the effectiveness and applicability of the DOLA framework to learning activities for STEM classes. After classification of overt learning activities as being active, constructive or interactive, I then tested the ICAP hypothesis, which states that student learning is more effective in interactive activities than constructive activities, which are more effective than active activities, which are more effective than passive activities. I conducted two studies (Study 1 and Study 2) to determine how and to what degree differentiated activities affected students' learning outcomes. For both studies, I measured students' knowledge of materials science and engineering concepts. Results for Study 1 showed that students scored higher on all post-class quiz questions after participating in interactive and constructive activities than after the active activities. However, student scores on more difficult, inference questions suggested that interactive activities provided significantly deeper learning than either constructive or active activities. Results for Study 2 showed that students' learning, in terms of gain scores, increased systematically from passive to active to constructive to interactive, as predicted by ICAP. All the increases, from condition to condition, were significant. Verbal analysis of the students' dialogue in interactive condition indicated a strong correlation between the co-construction of knowledge and learning gains. When the statements and responses of each student build upon those of the other, both students benefit from the collaboration. Also, the linear combination of discourse moves was significantly related to the adjusted gain scores with a very high correlation coefficient. Specifically, the elaborate type discourse moves were positively correlated with learning outcomes; whereas the accept type moves were negatively correlated with learning outcomes. Analyses of authentic activities in a STEM classroom showed that they fit within the taxonomy of the DOLA framework. The results of the two studies provided evidence to support the predictions of the ICAP hypothesis. / Dissertation/Thesis / Ph.D. Curriculum and Instruction 2012

A qualidade do ensino contábil tem sido questionada, tanto pelo resultado de exames de avaliação, quanto pelo próprio mercado de trabalho que recebe o profissional recém-formado. Cada vez mais vem sendo exigido do profissional contábil novos conhecimentos e habilidades Assim, argumenta-se que o processo de ensino precisa se modificar para oferecer o desenvolvimento de tais conhecimentos e habilidades. Dado este contexto, o objetivo geral desse estudo é entender as implicações que o uso da estratégia pedagógica \"Método do Caso\" traz para o ensino da contabilidade, à luz da Teoria da Aprendizagem Significativa. Como objetivos específicos, a tese busca i) analisar um caso de destaque de uma disciplina de graduação de contabilidade introdutória que utiliza o método do caso em Portugal; ii) entender a motivação dos docentes para implantar o método, bem como desafios e oportunidades durante a prática; iii) posicionar a visão do estudante nesse processo; e iv) traçar um paralelo com o ensino contábil no Brasil. Para alcançar o objetivo foi adotada a abordagem qualitativa de pesquisa e a estratégia de pesquisa estudo de caso único. A construção dos dados foi realizada por meio de análise documental, observação direta, entrevista semiestruturada e questionário aplicado aos alunos. A partir da análise dos dados construídos foi possível relacionar o método do caso de ensino com os pressupostos da aprendizagem significativa. A aprendizagem significativa se preocupa em entender como ocorre o processo de aprendizagem por uma ótica cognitiva, partindo da premissa de que ela ocorre quando existem conhecimentos prévios relevantes, o material é potencialmente significativo e o aluno está disposto a aprender de maneira significativa. Ao lançar mão do método do caso de ensino, o docente oferece a oportunidade de o aluno aprender de forma significativa, pois o conhecimento prévio relevante pode ser transmitido em aulas teóricas, o caso é um material potencialmente significativo, pois é construído com base em empresas reais e locais, e é, inerentemente, uma metodologia de aprendizagem ativa. No entendimento dos alunos, a disciplina Contabilidade Financeira I - o caso estudado - desempenha o papel de promover a aproximação entre teoria e prática. Para os docentes entrevistados, adotar o método do caso foi a solução encontrada para modernizar a disciplina e alterar o formato da mesma para uma perspectiva de articulação com os negócios, ao invés de uma visão tecnicista. Em um paralelo com o ensino contábil no Brasil, evidenciouse que as barreiras culturais não se apresentam como empecilho para a adoção da mesma dinâmica em uma disciplina semelhante. As análises apresentadas podem ser úteis para professores de contabilidade e coordenadores de curso que busquem alternativas para melhorar a qualidade do ensino e a efetividade do curso e a qualidade da formação dos futuros profissionais contábeis. / The quality of accounting education has been questioned, both by the results of evaluation exams and by the labor market itself that receives the newly trained professional. Increasingly, the accounting professional is required to have new knowledge and skills. Thus, the teaching method needs to be modified to offer the development of such knowledge and skills. Given this context, the general objective of this study is to understand the implications of the adoption of the pedagogical strategy \"Case Method\" can bring to the teaching of accounting in the light of Significant Learning Theory. As specific objectives, the thesis seeks to i) analyze a exemplar case of an undergraduate course that uses the case method at a prestigious institution in Portugal; ii) understand the motivation of teachers to implement the method, as well as challenges and opportunities during practice; iii) position the student\'s vision in this process; and iv) draw a parallel with accounting education in Brazil. To achieve the goal was adopted the qualitative research approach and the research strategy single case study. The construction of the data was done through documentary analysis, direct observation, semi-structured interview and questionnaire applied to the students. From the analysis of the constructed data, it was possible to relate the case method to the assumptions of meaningful learning. Meaningful learning is concerned with understanding how the learning process occurs from a cognitive point of view, starting from the premise that it occurs when there is relevant prior knowledge, the material is potentially significant, and the student is willing to learn in a meaningful way. By using the case method, the teacher offers the opportunity for the learner to learn meaningfully, since the relevant prior knowledge can be conveyed in theoretical lessons, the case is a potentially significant material as it is constructed based on real and local companies and the case method is inherently an active methodology. In the understanding of the students, the discipline Financial Accounting I - the case studied - plays the role of promoting the approximation between theory and practice. For the teachers interviewed, adopting the case method was the solution found to modernize the discipline and change the format of the same to a perspective of articulation with the business instead of a technicality vision. In parallel with accounting education in Brazil, it has been shown that cultural barriers do not present as an obstacle to the adoption of the same discipline format in similar accounting introductory courses. The analyzes presented may be useful for teachers and coordinators who seek alternatives to improve the quality of teaching and learning process and the effectiveness of the course and the quality of the formation gathered by future accounting professionals.

This thesis deals with the incorporation into teaching physics at the elementary school. It highlights the importance of conducting experiments on this subject. It deals with the role of improvised devices in physics education and their active use of the students. Examines and compares the benefits of active participation of pupils in the classroom for their motivation and better understanding and reinforce the learning.

Submitted by Vinicius Gomes de Lima (viniciuslimafgv@gmail.com) on 2017-05-19T18:15:35Z No. of bitstreams: 1 dissertacao_viniciuslima.pdf: 333215 bytes, checksum: db37308bb8e35440208daa9ef10ee9ba (MD5) / Approved for entry into archive by Suzinei Teles Garcia Garcia (suzinei.garcia@fgv.br) on 2017-05-19T18:36:54Z (GMT) No. of bitstreams: 1 dissertacao_viniciuslima.pdf: 333215 bytes, checksum: db37308bb8e35440208daa9ef10ee9ba (MD5) / Made available in DSpace on 2017-05-19T18:40:11Z (GMT). No. of bitstreams: 1 dissertacao_viniciuslima.pdf: 333215 bytes, checksum: db37308bb8e35440208daa9ef10ee9ba (MD5) Previous issue date: 2017-04-24 / This paper investigates peer effects in higher education in an environment of active learning that gives great importance for students’ interaction through group work. Our empirical strategy uses exogenous variation in group composition to estimate peer effects in different exercises. We find no evidence of peer effects in a basic linear-in-means specification considering all assigned peers. However, we find positive and statistically significant impact of peers coming from student’s same high school. We also find no evidence of peer effects with a model that takes into account student and peers’ position in the ability distribution. / Este trabalho investiga efeito de pares no ensino superior em um ambiente de active learning que dá grande importância à interação dos estudantes através do trabalho em grupo. A estratégia empírica utiliza variação exógena na composição dos grupos para estimar o efeito dos pares em diferentes exercícios. Não encontramos evidência de efeito de pares numa especificação linear-in-means básica considerando todos os pares do grupo atribuído ao aluno. Entretanto, encontramos efeito positivo e estatisticamente significante de pares que frequentaram a mesma escola de ensino médio. Não encontramos evidência de efeito de pares em um modelos que procura considerar a posição dos alunos na distribuição de habilidade.

Made available in DSpace on 2016-08-29T15:36:07Z (GMT). No. of bitstreams: 1 tese_9520_Dissertação - Vitor Jurtlero de Freitas (2)-versão final que foi para a biblioteca central.pdf: 5077183 bytes, checksum: 7e679664a025d1263ae11fa0a56791fc (MD5) Previous issue date: 2015-12-18 / CAPES / O método ensino da flipped classroom tem alcançado uma penetração cada vez maior nas mídias e salas de aula pelo mundo. Nos últimos 3 anos os jornais New York Times e Washington Post publicaram um total de 20 artigos relacionados ao tema. A Flipped Learning Network, comunidade online voltada para a divulgação, capacitação e integração de professores interessados na inversão da sala de aula possui mais de 20.000 membros e realizou em julho de 2015 sua oitava conferência anual de flipped, na Universidade Estadual de Michigan. A flipped classroom, ou sala de aula invertida, baseia-se na relocação do papel da palestra instrucional para fora do ambiente escolar, através da utilização de vídeo-aulas online, e na aplicação do dever de casa dentro da sala de aula por meio de atividades cooperativas e colaborativas baseadas na active learrning. O objetivo dessa inversão sala de aula x dever de casa é aproveitar o tempo que seria gasto em sala com a exposição do conteúdo para desenvolver a compreensão desse conteúdo através de atividades mediadas pelo professor. Esse trabalho relata a implementação da flipped classroom para o ensino de conceitos físicos em cinco turmas da primeira série do ensino médio de uma escola pública. Nele, desenvolveu-se um estudo de caso com o objetivo de compreender as percepções dos alunos e do professor frente a esse método de ensino. Recorreu-se para a coleta de dados a instrumentos qualitativos, como entrevistas semiestruturadas e o inquérito por questões abertas e a instrumentos quantitativos como o inquérito por questões fechadas com escala do tipo Likert e a categorização das repostas presentes nas atividades desenvolvidas. O estudo permitiu concluir que: (i) o método de ensino da flipped classroom se provou eficaz para promover motivação ao ensino de física; (ii) as redes sociais demonstraram-se eficientes como um ambiente virtual de aprendizagem; (iii) as atividades baseadas na active learning criaram um ambiente de colaboração para a construção do conhecimento; (iv) os alunos envolvidos neste estudo apresentaram uma dependência considerável da figura do professor para o desenvolvimento das atividades propostas. (v) o método de ensino levou o professor a assumir um papel de mediador na construção do conhecimento. / The flipped classroom is based on the relocation of the role of instructional lecture out of the school environment, through the use of online video classes, and the implementation of the "homework" in the classroom through cooperative and collaborative activities based on active learrning. The goal of this inversion "classroom vs. homework" is to take advantage of the time that would be spent in the room with the exposure of the content to develop the understanding that content through activities mediated by professor. This teaching method has achieved increasing penetration in the media and classrooms around the world. In the last 3 years the newspapers New York Times and the Washington Post published a total of 20 articles related to these topic. The Flipped Learning Network, online community dedicated to the dissemination, training and integration of teachers interested in flipping the classroom has more than 20,000 members and held in July 2015 its eighth annual Conference of flipped, at Michigan State University. This research reports the implementation of the flipped classroom for the teaching of physical concepts related to force and motion in five classes of first grade high school to a public school. Here, has developed a case study with the aim to check the applicability of this method of teaching for the physical education in classes of first grade high school. Resorted to collecting data qualitative instruments, such as semi-structured interviews and the inquiry open questions, and quantitative instruments as the investigation closed issues with Likert type scale and categorization of answers in the activities undertaken. The study made it possible to conclude that: (i) the flipped classroom teaching method proved effective to promote motivation in the physics teaching; (ii) social networks have demonstrated efficient as a virtual learning environment; (iii) activities based on active learning created a collaborative environment for the construction of knowledge; (iv) students involved in this study showed a significant dependence of the figure of the teacher for the development of the proposed activities; (v) the teaching method led the professor to assume a role of mediator in the construction of knowledge.

Made available in DSpace on 2016-06-02T20:02:52Z (GMT). No. of bitstreams: 1 5356.pdf: 3061501 bytes, checksum: f37384a71c7814daeeab23c1a85a8acb (MD5) Previous issue date: 2013-03-19 / Universidade Federal de Sao Carlos / This paper aims to present, as experience report, an active approach to teaching about the content of Acoustics by contextualized knowledge of music, from the design and construction of sound artifacts with low cost materials for high school students. The activities were conducted in a public school of the state of São Paulo in the city of Mogi Guaçu. The acoustic content were worked both quantitatively and qualitatively. Qualitative analysis was applied to the study of the properties of sound musical artifacts produced by students taking into account the aspects of production, propagation and amplification of sound. The quantitative phase of didactic activity was conducted using the software "Soundcard Scope V1.4" which allowed the harmonic analysis of the sound produced by these devices and captured by the sound card of a computer. The active approach we present has its theoretical foundations in education and social constructivist approach is inspired by Kuhn's epistemology. Process is guided by social contextualization of knowledge and consideration, the teacher, the students' spontaneous conceptions, as well as the cooperative work between them. The mediation of the teacher tried to always be located within the Zone of Proximal Development (ZPD) of the students and thus contributing to the formalization of scientific concepts worked. We have also added strategies such as the use of Information and Communication Technologies (ICT) and the approach of the relations between Science, Culture and Society. The material production of students through the sound artifacts that confected, and the monitoring of the production process of these signs indicate that the work has motivated students for learning physics, and also contributed to make it meaningful to them, some concepts of acoustics. / O presente trabalho tem por objetivo apresentar, como relato de experiência, uma abordagem de ensino ativo sobre o conteúdo de Acústica contextualizado por conhecimentos de música, a partir da elaboração e construção de artefatos sonoros, com materiais de baixo custo, por alunos do Ensino Médio. As atividades foram desenvolvidas numa escola pública da rede estadual de São Paulo localizada na cidade de Mogi Guaçu. Os conteúdos de acústica foram trabalhados de forma quantitativa e qualitativa. A análise qualitativa foi aplicada ao estudo das propriedades sonoras dos artefatos musicais produzidos pelos alunos se levando em conta os aspectos de produção, propagação e ampliação do som. A fase quantitativa da atividade didática foi conduzida com a utilização do software Soundcard Scope V1.4 que possibilitou a análise harmônica do som produzido por tais aparatos e captado pela placa de som de um computador. A abordagem ativa que apresentamos tem seus fundamentos teóricos no enfoque sócioconstrutivista do ensino e é inspirada pela epistemologia de Kuhn. Se pauta pelo processo de contextualização social dos saberes e na consideração, pelo professor, das concepções espontâneas dos alunos, assim como pelo trabalho cooperativo entre eles. A mediação do professor procurou estar sempre situada dentro da Zona de Desenvolvimento Proximal (ZDP) dos alunos e, portanto, contribuindo para a formalização dos conceitos científicos trabalhados. Também agregamos estratégias como a utilização de Tecnologias da Informação e Comunicação (TIC) e a abordagem das relações entre a Ciência, a Cultura e a Sociedade. A produção material dos alunos, por meio dos artefatos sonoros que confeccionaram, assim como o acompanhamento do processo de produção dos mesmos apresentam indícios de que o trabalho motivou os alunos para o aprendizado da Física e, também, contribuiu para tornar mais significativo à eles, alguns conceitos de acústica.