Datasets, features, learning, and models in visual recognition

Visual recognition is a fundamental research topic in computer vision. This dissertation explores datasets, features, learning, and models used for visual recognition. In order to train visual models and evaluate different recognition algorithms, this dissertation develops an approach to collect object image datasets on web pages using an analysis of text around the image and of image appearance. This method exploits established online knowledge resources (Wikipedia pages for text; Flickr and Caltech data sets for images). The resources provide rich text and object appearance information. This dissertation describes results on two datasets. The first is Berg’s collection of 10 animal categories; on this dataset, we significantly outperform previous approaches. On an additional set of 5 categories, experimental results show the effectiveness of the method. Images are represented as features for visual recognition. This dissertation introduces a text-based image feature and demonstrates that it consistently improves performance on hard object classification problems. The feature is built using an auxiliary dataset of images annotated with tags, downloaded from the Internet. Image tags are noisy. The method obtains the text features of an unannotated image from the tags of its k-nearest neighbors in this auxiliary collection. A visual classifier presented with an object viewed under novel circumstances (say, a new viewing direction) must rely on its visual examples. This text feature may not change, because the auxiliary dataset likely contains a similar picture. While the tags associated with images are noisy, they are more stable when appearance changes. The performance of this feature is tested using PASCAL VOC 2006 and 2007 datasets. This feature performs well; it consistently improves the performance of visual object classifiers, and is particularly effective when the training dataset is small. With more and more collected training data, computational cost becomes a bottleneck, especially when training sophisticated classifiers such as kernelized SVM. This dissertation proposes a fast training algorithm called Stochastic Intersection Kernel Machine (SIKMA). This proposed training method will be useful for many vision problems, as it can produce a kernel classifier that is more accurate than a linear classifier, and can be trained on tens of thousands of examples in two minutes. It processes training examples one by one in a sequence, so memory cost is no longer the bottleneck to process large scale datasets. This dissertation applies this approach to train classifiers of Flickr groups with many group training examples. The resulting Flickr group prediction scores can be used to measure image similarity between two images. Experimental results on the Corel dataset and a PASCAL VOC dataset show the learned Flickr features perform better on image matching, retrieval, and classification than conventional visual features. Visual models are usually trained to best separate positive and negative training examples. However, when recognizing a large number of object categories, there may not be enough training examples for most objects, due to the intrinsic long-tailed distribution of objects in the real world. This dissertation proposes an approach to use comparative object similarity. The key insight is that, given a set of object categories which are similar and a set of categories which are dissimilar, a good object model should respond more strongly to examples from similar categories than to examples from dissimilar categories. This dissertation develops a regularized kernel machine algorithm to use this category dependent similarity regularization. Experiments on hundreds of categories show that our method can make significant improvement for categories with few or even no positive examples.

A mixed logit model for predicting exit choice during building evacuations.

Knowledge on human behaviour in emergency is crucial to increase the safety of buildings and transportation systems. Decision making during evacuations implies different choices, of which one of the most important concerns the escape route. The choice of a route may involve local decisions between alternative exits from an enclosed environment. This work investigates the influence of environmental (presence of smoke, emergency lighting and distance of exit) and social factors (interaction with evacuees close to the exits and with those near the decision-maker) on local exit choice. This goal is pursued using an online stated preference survey carried out making use of non-immersive virtual reality. A sample of 1,503 participants is obtained and a Mixed Logit Model is calibrated using these data. The model shows that presence of smoke, emergency lighting, distance of exit, number of evacuees near the exits and the decision-maker, and flow of evacuees through the exits significantly affect local exit choice. Moreover, the model points out that decision making is affected by a high degree of behavioural uncertainty. Our findings support the improvement of evacuation models and the accuracy of their results, which can assist in designing and managing building and transportation systems. The main contribution of this work is to enrich the understanding of how local exit choices are made and how behavioural uncertainty affects these choices.

Tutoría en E-Learning. Funciones y roles del tutor en la formación online

La teleformación, de manera paralela a Internet, ha tenido una rápida y constante evolución a lo largo de los últimos años. Al mismo tiempo, dicha modalidad formativa ha ido cobrando mucha relevancia en el ámbito de la formación, en particular, en la capacitación de las personas trabajadoras. Teniendo en cuenta las ventajas que presenta la teleformación, se hace necesario que se realicen avances y mejoras de la calidad en el desarrollo de los procesos incluidos en ella, como es el caso de la tutorización. Por estas razones esta investigación se plantea: ¿Qué funciones realizan los tutores y tutoras de e-learning durante las acciones formativas en las que participan?, y ¿cómo perciben esas funciones los estudiantes? Este trabajo investiga las formas de tutorización que ayudan a un mejor desempeño por parte del teletutor, promoviendo procesos de enseñanza-aprendizaje que contribuyan a disminuir el índice de abandonos de estudiantes. La metodología seguida emplea técnicas cuantitativas y cualitativas buscando en cada momento, el método que mejor dé respuesta a los objetivos planteados inicialmente. Para la realización de la misma se han tomado dos muestras. Una de 707 estudiantes de cursos online pertenecientes a un proyecto de formación continua de trabajadores de pequeña y mediana empresas, y autónomos, La segunda muestra la conforman 8 tutores que han participado en el mencionado proyecto. A través de esta investigación se han podido analizar y estudiar las dimensiones y funciones que han realizado los tutores durante las distintas acciones formativas llevadas a cabo y se ha llegado a conocer las percepciones que tienen los estudiantes sobre las distintas dimensiones o funciones que ha empleado cada tutor en las mismas. Se consiguió identificar y comprender las tareas y funciones puestas en práctica por los tutores durante las acciones formativas llevadas a cabo, y las diferencias que se produjeron en ellas en cada uno de los docentes. Así mismo los resultados han permitido organizar diversos sistemas de categorías en otros tantos marcos “teóricos”, representados a través de diagramas comprensivos, que nos ayudan a entender de manera más adecuada las relaciones entre categorías y dimensiones de cada uno de los tutores del estudio. Ha sido posible incluso llegar a establecer tipologías de tutores según las funciones y roles desarrollados, los recursos más utilizados en su labor de tutorización y los rasgos que caracterizan a los estudiantes que realizan este tipo de formación.

Using interaction data for improving the offline and online evaluation of search engines

This thesis investigates how web search evaluation can be improved using historical interaction data. Modern search engines combine offline and online evaluation approaches in a sequence of steps that a tested change needs to pass through to be accepted as an improvement and subsequently deployed. We refer to such a sequence of steps as an evaluation pipeline. In this thesis, we consider the evaluation pipeline to contain three sequential steps: an offline evaluation step, an online evaluation scheduling step, and an online evaluation step. In this thesis we show that historical user interaction data can aid in improving the accuracy or efficiency of each of the steps of the web search evaluation pipeline. As a result of these improvements, the overall efficiency of the entire evaluation pipeline is increased. Firstly, we investigate how user interaction data can be used to build accurate offline evaluation methods for query auto-completion mechanisms. We propose a family of offline evaluation metrics for query auto-completion that represents the effort the user has to spend in order to submit their query. The parameters of our proposed metrics are trained against a set of user interactions recorded in the search engine’s query logs. From our experimental study, we observe that our proposed metrics are significantly more correlated with an online user satisfaction indicator than the metrics proposed in the existing literature. Hence, fewer changes will pass the offline evaluation step to be rejected after the online evaluation step. As a result, this would allow us to achieve a higher efficiency of the entire evaluation pipeline. Secondly, we state the problem of the optimised scheduling of online experiments. We tackle this problem by considering a greedy scheduler that prioritises the evaluation queue according to the predicted likelihood of success of a particular experiment. This predictor is trained on a set of online experiments, and uses a diverse set of features to represent an online experiment. Our study demonstrates that a higher number of successful experiments per unit of time can be achieved by deploying such a scheduler on the second step of the evaluation pipeline. Consequently, we argue that the efficiency of the evaluation pipeline can be increased. Next, to improve the efficiency of the online evaluation step, we propose the Generalised Team Draft interleaving framework. Generalised Team Draft considers both the interleaving policy (how often a particular combination of results is shown) and click scoring (how important each click is) as parameters in a data-driven optimisation of the interleaving sensitivity. Further, Generalised Team Draft is applicable beyond domains with a list-based representation of results, i.e. in domains with a grid-based representation, such as image search. Our study using datasets of interleaving experiments performed both in document and image search domains demonstrates that Generalised Team Draft achieves the highest sensitivity. A higher sensitivity indicates that the interleaving experiments can be deployed for a shorter period of time or use a smaller sample of users. Importantly, Generalised Team Draft optimises the interleaving parameters w.r.t. historical interaction data recorded in the interleaving experiments. Finally, we propose to apply the sequential testing methods to reduce the mean deployment time for the interleaving experiments. We adapt two sequential tests for the interleaving experimentation. We demonstrate that one can achieve a significant decrease in experiment duration by using such sequential testing methods. The highest efficiency is achieved by the sequential tests that adjust their stopping thresholds using historical interaction data recorded in diagnostic experiments. Our further experimental study demonstrates that cumulative gains in the online experimentation efficiency can be achieved by combining the interleaving sensitivity optimisation approaches, including Generalised Team Draft, and the sequential testing approaches. Overall, the central contributions of this thesis are the proposed approaches to improve the accuracy or efficiency of the steps of the evaluation pipeline: the offline evaluation frameworks for the query auto-completion, an approach for the optimised scheduling of online experiments, a general framework for the efficient online interleaving evaluation, and a sequential testing approach for the online search evaluation. The experiments in this thesis are based on massive real-life datasets obtained from Yandex, a leading commercial search engine. These experiments demonstrate the potential of the proposed approaches to improve the efficiency of the evaluation pipeline.

Learning Binary Code Representations for Effective and Efficient Image Retrieval

The size of online image datasets is constantly increasing. Considering an image dataset with millions of images, image retrieval becomes a seemingly intractable problem for exhaustive similarity search algorithms. Hashing methods, which encodes high-dimensional descriptors into compact binary strings, have become very popular because of their high efficiency in search and storage capacity. In the first part, we propose a multimodal retrieval method based on latent feature models. The procedure consists of a nonparametric Bayesian framework for learning underlying semantically meaningful abstract features in a multimodal dataset, a probabilistic retrieval model that allows cross-modal queries and an extension model for relevance feedback. In the second part, we focus on supervised hashing with kernels. We describe a flexible hashing procedure that treats binary codes and pairwise semantic similarity as latent and observed variables, respectively, in a probabilistic model based on Gaussian processes for binary classification. We present a scalable inference algorithm with the sparse pseudo-input Gaussian process (SPGP) model and distributed computing. In the last part, we define an incremental hashing strategy for dynamic databases where new images are added to the databases frequently. The method is based on a two-stage classification framework using binary and multi-class SVMs. The proposed method also enforces balance in binary codes by an imbalance penalty to obtain higher quality binary codes. We learn hash functions by an efficient algorithm where the NP-hard problem of finding optimal binary codes is solved via cyclic coordinate descent and SVMs are trained in a parallelized incremental manner. For modifications like adding images from an unseen class, we propose an incremental procedure for effective and efficient updates to the previous hash functions. Experiments on three large-scale image datasets demonstrate that the incremental strategy is capable of efficiently updating hash functions to the same retrieval performance as hashing from scratch.

The interaction factor as a challenge for foreign language learning in the technology based education: a pedagogical reflection

The author carries out a pedagogical reflection on how the technology driven distance learning repeatedly neglects the scientific achievements of Second Language Acquisition and Language Pedagogy. Seeing communicative competence as a major goal of a language classroom, she presents the main challenges that the communicative approach poses to distance learning. To this end, a general distance learning theory by Moore is adapted to the needs of language education, through a distinction between three aspects of learner interaction – with the teacher, with other learners and with content. In this three-dimensional paradigm the learner is seen as the main actor of the process, the teacher as a facilitator, the text as a main source of communicative data and the learner autonomy as the fundament of the process.

Model Probability in Self-organising Maps

Growing models have been widely used for clustering or topology learning. Traditionally these models work on stationary environments, grow incrementally and adapt their nodes to a given distribution based on global parameters. In this paper, we present an enhanced unsupervised self-organising network for the modelling of visual objects. We first develop a framework for building non-rigid shapes using the growth mechanism of the self-organising maps, and then we define an optimal number of nodes without overfitting or underfitting the network based on the knowledge obtained from information-theoretic considerations. We present experimental results for hands and we quantitatively evaluate the matching capabilities of the proposed method with the topographic product.

A satisfação e a intenção de continuidade de uso em servços de e-learning:validação empírica de um modelo aplicado no serviço público

The present study aims to investigate the constructs of Technological Readiness Index (TRI) and the Expectancy Disconfirmation Theory (EDT) as determinants of satisfaction and continuance intention use in e-learning services. Is proposed a theoretical model that seeks to measure the phenomenon suited to the needs of public organizations that offer distance learning course with the use of virtual platforms for employees. The research was conducted from a quantitative analytical approach, via online survey in a sample of 343 employees of 2 public organizations in RN who have had e-learning experience. The strategy of data analysis used multivariate analysis techniques, including structural equation modeling (SEM), operationalized by AMOS© software. The results showed that quality, quality disconfirmation, value and value disconfirmation positively impact on satisfaction, as well as disconfirmation usability, innovativeness and optimism. Likewise, satisfaction proved to be decisive for the purpose of continuance intention use. In addition, technological readiness and performance are strongly related. Based on the structural model found by the study, public organizations can implement e-learning services for employees focusing on improving learning and improving skills practiced in the organizational environment

The interaction of economic and pedagogical ideals in the context of workplace learning in Germany: a framework for empirical research – inspired by business ethics

As an effect of marketisation, the importance of workplace learning in Germany has increased. The article follows up on the long-standing discourse around the question of how economic and pedagogical ideals interact in this context. In order to develop a theoretical framework for empirical research, three major positions of the discipline of business ethics are introduced. Business ethics in more abstract ways deals with the very same question, namely how do ideas such as profit orientation interact with other norms and values? The new perspectives show that the discourse has been hitherto based on a specific understanding of economy. In order to derive an empirical answer to the research question, the question is re-formulated as follows: Which values are inherent in the decisions taken? Consequently, it suggests using the concept of ‘rationalities of justification’ for empirical research. The article shows how this concept can be applied by conducting a test run. (DIPF/Orig.)

Community School of São Miguel de Machede: A decade of Learning

the Community School of São Miguel de Machede exists since 1998. A model of Community Education has been developed in this decade of existence, which not being confined to the frequent profiles of the most common approaches in Adult Education, has been the result of a process of symbiosis between a practice that normally precedes the conceptualization and a thought which has always expressed the concern of interpreting and enrich that practice. Setting on a model of learning based on the PADéCA – Program of Helping the Development of the Capacity to Learn, proposed by Berbaum (1988), the Community School of São Miguel de Machede has been developing several activities centred on a fundamental concern: to create easy and qualified accesses, in this community (council of Evora), so that the respective members can learn to exercise their principal rights of citizenship, in the territory where they live and in a circumstance of equality of opportunities in relation to the remaining fellow countrymen. Being a project with a decade of life, it is now possible to speak of a history full of stories and learning experiences, which occurred as a result of a rich interaction between the initial thoughts and impulses of the theoretical approaches and a reality full of unexpectedness, mutability and humanity resulting from the complexity that a living community presents, with a history and a present, but not always with clear and positive idea about the respective future.

Curso online para a parte pral do Celpe-Bras : contribuições da avaliação de proficiência para o ensino-aprendizagem de PLE

Dissertação (mestrado)—Universidade de Brasília, Departamento de Línguas Estrangeiras e Tradução, Programa de Pós-Graduação em Linguística Aplicada, 2016.

Subjetividade e interação nos fóruns online reflexões sobre a permanência em educação a distância

Este artigo foi desenvolvido com objetivo de produzir indicadores que possam melhorar a funcionalidade dos fóruns online e contribuir numa maior permanência dos estudantes da Educação a Distância. Foi realizada uma análise, orientada pela Epistemologia Qualitativa, dos processos subjetivos e interacionais produzidos nos fóruns de apresentação e fóruns temáticos de duas disciplinas de formação pedagógica – (1) Estratégias de Ensino e Aprendizagem e (2) A Psicologia e a Construção do Conhecimento – ofertadas nos cursos de Licenciatura em Teatro, Música e Artes Visuais, UAB/UnB. As informações produzidas apontam para a necessidade de reconhecimento e valorização do estudante como sujeito na aprendizagem, a consolidação da presença pedagógica do tutor, a valorização dos fóruns como espaços de aprendizagem e a produção de espaços sociais de pertencimento. ______________________________________________________________________________ ABSTRACT

Online Resources for Improving EFL Teaching and Learning

El presente artículo plantea el beneficio de utilizar recursos en línea en la enseñanza del inglés. Destaca que los estudiantes fortalecerán, no solo el uso de la lengua meta, sino también el de los recursos tecnológicos. Para ejemplificar, se presenta una serie de ejercicios en línea que desarrollan diversas habilidades de la lengua meta como también las ventajas y desventajas del uso de los mismos. Por último, se comparten los resultados obtenidos de una encuesta aplicada sobre el uso de recursos en línea en las clases de inglés.The benefit of using online sources in the EFL class is analyzed here starting from the perspective that this helps students improve not only their use of the language but also their use of technology. Sample online exercises focusing on the development of different language skills are described here, along with the advantages and disadvantages of using online sources. Finally, the results obtained from a survey on the use of online sources in the EFL classes are presented.

Real-time audio interaction in serious games for music learning

In this LBD, we present several Apps for playing while learning music or for learning music while playing. The core of all the games is based on the good performance of the real-time audio interaction algorithms developed by the ATIC group at Universidad de Ma ́laga (SPAIN).