Dissertations / Theses on the topic 'Examprep'

The problem of simulating a crowd to see how it would behave in certain situations or just to create a realistic-looking scene to be used in a movie or video game is important and complex, and there are many different methods to solve it. This project is primarily an investigation of the example-based crowd simulation method described in the article "Crowds by Example" by Lerner et al. In the article, traced video footage of crowds are used to create a data set. The simulation program continuously finds situations in the data set that resembles the current situations in the simulation and updates the simulation thereby. We implemented this for around 10 agents using Unity 3D. Example-based crowd simulations does not only, like some other types of crowd simulation methods (for example ORCA), take collision avoidance into account but also the more complex ways the human mind thinks and therefore does not always behave as one would predict. The main conclusion is that this method of simulating crowds has the potential to create more realistic simulations than other forms of crowd simulations. The downsides are that the time the program spends creating simulations can quickly get very high and to make realistic simulations a lot of video footage must be filmed and then traced.

Math & Science Education
Ph.D.
Researchers and instructors have only recently embraced the role of errors as vehicles for learning in the algebra classroom. Studying a mixture of correct and incorrect worked examples has been shown to be beneficial relative to correct worked examples alone. This study examines the effectiveness of having students generate, or anticipate, errors another student might make. Five Algebra 1 sections at a suburban mid-Atlantic public high school participated amid an early equation-solving unit. During teacher-led instruction, all five sections examined 2-3 correct worked examples. The final example varied across conditions. One section received an additional correct worked example. Two sections examined an incorrect worked example. The remaining two sections engaged in an error anticipation exercise where the teacher wrote an equation on the board and asked the students to predict errors another student might make in solving. The study measured conceptual and procedural knowledge, encoding ability, and student-generated errors. Although no meaningful significant differences were found, students in the error anticipation condition saw no difference in performance in conceptual and procedural items versus those who examined incorrect worked examples. Analysis that combined the error anticipation and incorrect worked examples conditions showed that those students trended toward outperforming those who examined correct examples only on procedural items. These results support further examination of error anticipation as a worthwhile instructional activity.
Temple University--Theses

Mislabeled examples affect the performance of supervised learning algorithms. Two novel approaches to this problem are presented in this Thesis. Both methods build on the hypothesis that the large margin and the soft margin principles of support vector machines provide the characteristics to select mislabeled examples. Extensive experimental results on several datasets support this hypothesis. The support vectors of the one-class and two-class SVM classifiers captures around 85% and 99% of the randomly generated label noise examples (10% of the training data) on two character recognition datasets. The numbers of examples that need to be reviewed can be reduced by creating a two-class SVM classifier with the non-support vector examples, and then by only reviewing the support vector examples based on their classification score from the classifier. Experimental results on four datasets show that this method removes around 95% of the mislabeled examples by reviewing only around about 14% of the training data. The parameter independence of this method is also verified through the experiments. All the experimental results show that most of the label noise examples can be removed by (re-)examining the selective support vector examples. This property can be very useful while building large labeled datasets.

"Submitted in partial fulfillment of the requirements for the degree of Master of Business Administration from the Naval Postgraduate School, June 2009."
Advisor(s): Euske, Kenneth ; Mutty, John. "June 2009." "MBA professional report"--Cover. Description based on title screen as viewed on July 14, 2009. DTIC Identifiers: LCU (Landing Craft Units). Author(s) subject terms: Budget, Model, Requirements-based, LCU, ACU-1, 1C6C, NBG-1, OPTAR, Operational Availability Includes bibliographical references (p. 73-76). Also available in print.

The use of examples is known to be important in learning, they should be “exemplary” and function as role-models.

Teaching and learning problem solving and programming in the object oriented paradigm is recognised as difficult. Object orientation is designed to handle complexity and large systems, and not with education in focus. The fact that object orientation often is used as first paradigm makes the design of examples even more difficult and important.

In this thesis, a survey of the literature is made to establish a set of characteristics for object orientation in general. This set of characteristics is then applied to the educational setting of introducing novices to object oriented problem solving and programming, resulting in a number of heuristics for educational purposes, called He[d]uristics. The proposed He[d]uristics are targeted towards educators designing small-scale examples for novices, and is an attempt to provide help in designing suitable examples, not a catalogue of good ones.

The He[d]uristics are discussed and exemplified and also evaluated versus the derived set of characteristics and known common problems experienced by novices.

Empirical studies have shown that learning from worked examples is an effective learning strategy. A worked example provides step-by-step explanations of how a problem is solved. Many studies have compared learning from examples to unsupported problem solving, and suggested presenting worked examples to students in the initial stages of learning, followed by problem solving once students have acquired enough knowledge. Recently, researchers have started comparing learning from examples to supported problem solving in Intelligent Tutoring Systems (ITSs). ITSs provide multiple levels of assistance to students, adaptive feedback being one of them. The goal of this research is to investigate using examples in constraint-based tutors by adding examples into SQL-Tutor. SQL-Tutor is a constraint-based tutor that teaches the Structured Query Language (SQL). Students with different prior knowledge benefit differently from studying examples; thus, another goal of the research is to propose an adaptive model that considers the student’s prior knowledge for providing worked examples. Evaluation of this research produced promising results. First, a fixed sequence of alternating examples and problems was compared with problems only and examples only. The result shows that alternating examples and problems is superior to the other two conditions. Then, a study was conducted, in which a fixed sequence of alternating worked examples and tutored problem solving is compared with a strategy that adapts the assistance level to students’ needs. The adaptive strategy determines the type of the task (a worked example, a faded example or a problem to be solved) based on how much assistance the student received in the previous problem. The results show that students in the adaptive condition learnt significantly more than their peers who were presented with the fixed sequence of worked examples and problem solving. The final study employed eye tracking and demonstrated that novices and advanced students study SQL examples differently. Such information can be used to provide proactive rather than reactive feedback messages to students’ actions.

This study investigated the usefulness of providing successful and unsuccessful example solutions in enhancing students' technological problem-solving efficiency. Prior research exploring worked example solutions indicated improved problem-solving efficiency when solutions were structured in a fashion that decreased the amount of extraneous cognitive load and increased the amount of germane cognitive load as specified by cognitive load theory. Fifty-one 7th and 8th grade students enrolled in technology education courses were selected from one school in the southwest region of Virginia. Participants completed three technological problem-solving tasks that included elevated load, cantilevered weight, and energy absorption using supply kits containing simple modeling materials. Problem-solving efficiency was determined by combining the amount of elapsed time across all three tasks. A 3 x 3 mixed factorial ANOVA was used to analyze the data. Data analysis revealed trends similar to worked example research in mathematics and science, but no significant difference among the three groups was found in this study.
Ph. D.

Currently, there is an urgent need to conduct analysis of teaching statistical courses at the Faculty of Management University of Economics in Prague in order to analyze the current teaching methods, which are at the time of this thesis in the current academic year 2009-2010. Another aim is to propose a new method of teaching that reflect emerging trends in education, which would help students to understand and absorb the learning.

Thesis (M.S. in Operations Research)--Naval Postgraduate School, March 2010.
Thesis Advisor: Alderson, David L. Second Reader: Church, Richard L. "March 2010." Author(s) subject terms: Evacuation, Network, Flow Optimization, Space-Time network. Includes bibliographical references (p. 41-42). Also available in print.

The present study combined subgoal learning and self-explanation frameworks to improve problem solving performance. Subgoal learning has been used to promote retention and transfer in procedural domains, such as programming. The primary method for learning subgoals, however, has been through passive learning methods, and passive learning methods are typically less effective than constructive learning methods. To promote constructive methods of learning subgoals, a subgoal learning framework was used to guide self-explanation. Self-explanation is an effective method for engaging learners to make sense of new information based on prior knowledge and logical reasoning. Self-explanation is typically more effective when learners receive some guidance, especially if they are novices, because it helps them to focus their attention on relevant information. In the present study, only some of the constructive learning methods produced better problem solving performance than passive learning methods. Learners performed best when they learned constructively and either received hints about the subgoals of the procedure or received feedback on the self-explanations that they constructed, but not when they received both hints and feedback. When students received both types of guidance, they did not perform better than those who learned subgoals through passive learning methods. These findings suggest that constructive learning of subgoals can further improve the benefits of learning subgoals, but there is an optimal level of guidance for students engaging in constructive learning. Providing too much guidance can be as detrimental as providing too little. This nuance is important for educators who engage their students in constructive learning and self-explanation to recognize and promote the best results.

The purpose of this study was to determine the effects of a simulation with workedexamples on the creation of episodic memories and the troubleshooting ability of maintenance technician students. Previous research shows that domain knowledge, conceptual knowledge, strategic knowledge, and episodic memories are all required to successfully troubleshoot. While domain, conceptual, and strategic knowledge can all be taught using traditional instruction, episodic memories require students to experience the actual troubleshooting of a fault. Simulations and worked examples are two instructional methods that have proven effective at teaching troubleshooting. This research specifically examined how a simulation combined with worked examples would affect 1) immediate troubleshooting abilities, 2) the creation of episodic memories, and 3) delayed troubleshooting abilities. This study was conducted in two stages and administered via a learning management system due to COVID-19 restrictions. The first stage included a pre-test, a training session using the simulation with worked examples, and an immediate post-test for near and far transfer of troubleshooting abilities. The second stage occurred one week later and included the final posttest for near and far transfer of troubleshooting abilities and creation of episodic memories. Answers to four troubleshooting questions on each of the pre-test and immediate and delayed post-tests were collected to determine any differences in the immediate and retained troubleshooting abilities. Answers to the solution mapping questions were collected to determine the creation of episodic memories. A repeated measure analysis of variance was conducted in SPSS to analyze the results of the troubleshooting pre- and post-tests. A correlational coefficient was used to determine any interaction between episodic memories and delayed troubleshooting abilities. Previous experience levels and participants’ major of study were also examined to determine their effect on the results. The findings show the simulation with worked examples had a statistically significant effect on delayed troubleshooting abilities and the created episodic memories had a positive correlation with the delayed troubleshooting, both with a medium effect size. However, the simulation with worked examples had no statistically significant effect on immediate troubleshooting abilities. Levels of previous experience and participants’ major of study had little effect on the results.

Informed parent advocates are essential to planning the educational outcomes of their children with special needs in the K through 12 public school system. However, inappropriate instructional techniques used in advocacy training may reduce trainees' learning outcomes by adding complexity and increasing cognitive load. This study examined whether using worked examples to break down complex problems into component parts to build long term schema could lower cognitive load and thus improve learning outcomes for parent advocacy trainees. Based on cognitive load theory, this 2 x 3 factorial design study examined the efficacy of noninteractive video lessons for parent trainees using worked examples to reduce extraneous cognitive load. Research questions explored the relationships between the independent variables of using worked examples and parents' perceived class relevance on the dependent variable, change in cognitive load of parent trainees, as well as the interaction between training type and perceived class relevance. Two groups of 65 adults in advocacy training (N = 130) participated in a video lesson in either the worked examples or nonworked examples format as part of their advocacy training. The NASA Task Load Index and the Perceived Class Relevance Survey instruments were used to measure cognitive load of trainees and perceptions of training relevance. Key findings included a significant main effect between the use of worked examples and change in cognitive load and significant interaction effects with the perception of class relevance. Training was developed for advocacy trainers in the use of worked examples for learners new to a domain. Implications for social change include improved learning outcomes for parents who must learn IEP terminology in beginning classes to effectively advocate for their children.

The success of deep learning is inseparable from the support of massive data. The vast amount of high-quality data is one of the most crucial prerequisites to train a robust deep learning model. However, the raw data collected in the real world usually has some defects which prevent their direct usage for training. In this thesis, we propose to employ the generative adversarial learning framework to generate high-quality data for training robust deep learning models. We focus on the generation of two of the most common data types: time series and images. For time series, we propose an adversarial recurrent time series imputation model to reconstruct the incomplete time series. Specifically, our model modifies the traditional Recurrent Neural Network (RNN) architecture to better capture the temporal dependencies and feature correlations and legitimately combines them to impute the missing part. Besides, we employ an element-wise generative adversarial learning framework to train the modified recurrent structure to generate more realistic data. Experiments on several real-world time series datasets demonstrate encouraging improvement of our model on the imputation performance as well as the classification accuracy. For image classification, recent studies find that deep learning models trained with natural images can be vulnerable to a kind of adversarially generated perturbations, namely adversarial perturbations. In this work, we improve the adversarial example generation and traditional AT framework from four aspects: adaptive perturbation size, diversified adversarial examples, stabilizing AT with massive contrastive adversaries, and adversarial robustness through representation disentanglement. The models presented above all demonstrate empirical remarkable performance improvement in terms of the quality of adversarial examples and model robustness.

Induction of classifiers from sets of preclassified training examples is one of the most popular machine learning tasks. This dissertation focuses on the techniques needed in the field of automated text categorization. Here, each document can be labeled with more than one class, sometimes with many classes. Moreover, the classes are hierarchically organized, the mutual relations being typically expressed in terms of a generalization tree. Both aspects (multi-label classification and hierarchically organized classes) have so far received inadequate attention. Existing literature work largely assumes that it is enough to induce a separate binary classifier for each class, and the question of class hierarchy is rarely addressed. This, however, ignores some serious problems. For one thing, induction of thousands of classifiers from hundreds of thousands of examples described by tens of thousands of features (a common case in automated text categorization) incurs prohibitive computational costs---even a single binary classifier in domains of this kind often takes hours, even days, to induce. For another, the circumstance that the classes are hierarchically organized affects the way we view the classification performance of the induced classifiers. The presented work proposes a technique referred to by the acronym "H-kNN-plus." The technique combines support vector machines and nearest neighbor classifiers with the intention to capitalize on the strengths of both. As for performance evaluation, a variety of measures have been used to evaluate hierarchical classifiers, including the standard non-hierarchical criteria that assign the same weight to different types of error. The author proposes a performance measure that overcomes some of their weaknesses. The dissertation begins with a study of (non-hierarchical) multi-label classification. One of the reasons for the poor performance of earlier techniques is the class-imbalance problem---a small number of positive examples being outnumbered by a great many negative examples. Another difficulty is that each of the classes tends to be characterized by a different set of characteristic features. This means that most of the binary classifiers are induced from examples described by predominantly irrelevant features. Addressing these weaknesses by majority-class undersampling and feature selection, the proposed technique significantly improves the overall classification performance. Even more challenging is the issue of hierarchical classification. Here, the dissertation introduces a new induction mechanism, H-kNN-plus, and subjects it to extensive experiments with two real-world datasets. The results indicate its superiority, in these domains, over earlier work in terms of prediction performance as well as computational costs.

Powerful classifiers as neural networks have long been used to recognise images; these images might depict objects like animals, people or plain text. Distorted images affect the neural network's ability to recognise them, they might be distorted or changed due to distortions related to the camera.Camera related distortions, and how they affect the accuracy, have previously been explored. Recently, it has been proven that images can be intentionally made harder to recognise, an effect that last even after they have been photographed.Such images are known as adversarial examples.The purpose of this thesis is to evaluate how well a neural network can recognise adversarial examples which are also distorted. To evaluate the network, the adversarial examples are distorted in different ways and thereafter fed to the neural network.Different kinds of distortions (rotation, blur, contrast and skew) were used to distort the examples. For each type and strength of distortion the network's ability to classify was measured.Here, it is shown that all distortions influenced the neural network's ability to recognise images.It is concluded that the type and strength of a distortion are important factors when classifying distorted adversarial examples, but also that some distortions, rotation and skew, are able to keep their characteristic influence on the accuracy, even if they are influenced by other distortions.

School Psychology
Ph.D.
This study was designed to compare the effects of the use of worked examples and self-explanation on motivation for male and female students. More specifically, the present study examines whether there are differences between males and females with regards to their achievement goals and if gender plays a role in how students respond to questions about their motivation in the presence of other male or female students. Comparisons of student responses on Achievement Goal Questionnaire-Revised (AGQ-R) and the Patterns of Adaptive Learning Scales (PALS) were also conducted. Participants were 147 seventh-, eighth- and ninth-grade non-honors Algebra I students (82 girls and 65 boys) from three schools and eight classrooms within the same school district on the east cost of the United States of America. Results replicated the finding that females have more mastery goals than males, however no gender differences were found for either performance. In addition, it appears that students respond differently to some questions about their motivation in the presence of other male or female students. Finally, the AGQ-R and the PALS appear to be consistent representations of students' achievement goals.
Temple University--Theses

Syftet med studien var att undersöka priming med exempels påverkan på skapelsers kreativitet. Effekterna av priming skulle testas i formen av egenskapspriming såväl som konceptuell priming och mätas utifrån både en användbarhetsaspekt och en originalitetsaspekt. Ett delsyfte var också att undersöka sambandet mellan intresset för att ta fram en skapelse och en skapelses grad av originalitet samt användbarhet. Primingeffekterna prövades genom ett experiment där 36 studenter delades in i 3 grupper som fick i uppgift att ta fram IT-lösningar för kollektivtrafiken. 2 av grupperna primades med var sitt exempel på hög originalitet respektive hög användbarhet. Effekterna av egenskapspriming undersöktes sedan genom en jämförelse mellan grupperna över hur stor andel av de framtagna lösningarna som innehöll egenskaper från exemplen. För att undersöka effekterna av konceptuell primingen gjordes en jämförelse av originaliteten och användbarhet hos dessa gruppers lösningar jämfört med de framtagna av en kontrollgrupp utan tillgång till exempel. Lösningarnas originalitet och användbarhet bedömdes av en expertgrupp bestående av forskare inom kreativitet och branschaktiva inom området. Denna expertgrupp bedömde även sitt intresse för att ta fram de olika lösningarna. Denna bedömning tillsammans med de två tidigare användes för att undersöka sambandet mellan intresset för en lösning och lösningens originalitet respektive användbarhet. Resultatet av studien visade på ett signifikant positivt samband hos båda grupperna av bedömare mellan en lösnings användbarhet och intresset för att ta fram lösningen. Ett signifikant positivt samband gällande en lösnings originalitet och intresset för denne kunde dock endast uppvisas hos forskare medan de branschaktiva visade på ett osignifikant negativt samband mellan de två. Resultatet av egenskapspriming och konceptuell priming visade inte överlag på några signifikanta skillnader mellan grupperna. Studien anses vara av vikt då den visar på hur en ökning av kreativiteten på skapelser inom en organisation inte behöver betyda en ökning av organisationens kreativitet p.g.a. ointresset i att omfamna och satsa på dessa lösningar. Vidare visar studien på hur primingeffekter är svåra att skapa i situationer som i högre grad ska efterlikna de i verkliga livet och hur vidare forskning behövs på området för att kartlägga variabler som påverkar priming och primings begränsningar.
The aim of the study was to examine priming with example’s impact on ideas creativity, regarding usability and novelty both as a result of targeting priming and conceptual priming. A subsidiary aim was to examine the relationship between an idea’s degree of originality and usefulness and the interest in pursuing the idea. The effects of priming were tried in an experiment involving 36 students who were asked to develop IT solutions for public transport. The students were randomly divided into two experimental groups and one control group. The members of the experimental group were primed with an example with high originality respectively usability. The solutions were given two indexes on how many of the properties of respective examples were found in the solutions. Their degree of originality and usefulness were also assessed by a group of experts. This group also assessed their interest in the respective solution. The study revealed a significant positive correlation between a solution's usability and interest of the solution. A positive significant correlation could also be seen between a solution's originality and the interest of the solution, but only in the group of analysts consisting of scientists. Among the industry-experts the results instead pointed toward a negative relationship between originality and interest, although a non-significant relationship. There were no significant differences between the group’s solutions regarding their degree of originality and usefulness nor the frequency in which features from the example was found in the solutions. The study is of importance because it shows how increases in the creativity of solutions produced by individuals and employees not necessarily lead to an increase in the creativity of an organization, due to lack of interest in embracing and invest in these creative solutions. The study also reflects the difficulties in transferring priming effects discovered in a lab environment to situations that more closely resembles those in real life. Further research is necessary to broaden the knowledge regarding variables that affects priming and the limits between which priming exist.

The aim of the thesis is to gain knowledge concerning what pupils aged 8 and 9 need to learn to become familiar with negative numbers. The framework used in this research, variation theory, impliesthat students' problems in learning what was intended may have to do with the fact that some critical aspects of the studied object have not yet been discerned by the student. To get the pupils to understand the idea behind each critical aspect, carefully constructed examples were used. According to variation theory it is necessary to experience differences before you experience similarities. To answer the research question data was collected by using the learning study model. It is characterized by an iterative design where I as a researcher collaborate with teachers to try to find and orchestrate the critical aspects. The method is interventionist, which means that interventions are done in teaching. In the learning study I have cooperated with two primary school teachers and 64 pupils in four different classes. The data consists of video-recordings of lessons, pre- and posttests, interviews with pupils and notes from the meetings of the learning study group. When planning lessons as well as analyzing data, concepts relating to the theory of variation have been used as analytical tools. This thesis contributes to research by investigating in detail what aspects students need to differentiate in order to become familiar with negative numbers. The results show that the pupils needed not only to discern, but also to differentiate three different critical aspects: To differentiate the values of two negative numbers. To differentiate the function of the minuend versus the function of the subtrahend in a subtraction. To differentiate the minus sign for negative numbers versus the minus sign for subtraction.

Learning from worked examples typically involves study activities, involving reading such materials carefully and attempting to understand the information presented. Considerable evidence has amassed regarding the benefits for novices of studying appropriately constructed worked examples paired with practice questions. However, prior research from the cognitive and sports psychology literatures suggests mental practice of worked examples may be an effective adjunct to studying such materials. Meta-analyses of these literatures suggest the utility of mental practice depends upon the degree of cognitive elements contained within a task, and also suggest that some prior knowledge of a task is necessary for mental practice to be effective. The present series of studies aimed to identify conditions under which mental practice is effective in educationally realistic, highly cognitive domains. Based on the above meta-analytic results, mental practice was hypothesised to enhance learning over further study in highly cognitive domains, but only when students either had sufficient prior knowledge, or were able to develop such knowledge over the course of an instructional intervention. Study activities were primarily expected to support knowledge acquisition, while imagining-based activities (mental practice) were expected to support knowledge automation. The experiments herein thus investigated interactions between levels of prior knowledge, complexity of instructional material, and levels of learning from imagination versus conventional study strategies. In Experiment 1, under conditions of low prior knowledge and complex material (HTML), students who studied worked examples outperformed those who imagined. Experiment 2, using simplified but still complex materials and a similar participant pool, found no differences between conditions, but Experiment 3, using stricter experimental design, found a study effect. In Experiment 4, an imagination effect was found under high prior knowledge. Experiment 5, using less experienced learners, suggested those who studied outperformed those whom imagined on acquisition questions. Experiment 6 found a sequence of study then imagination is more effective than imagination then study. The results have broad application for effective sequencing of these instructional strategies, but development of an accurate metric for imagination "readiness" is required to advance theory and practice, and more evidence is needed for a schema automation explanation of mental practice effects.

Bayesian statistical methods have long been computationally out of reach because the analysis often requires integration of high-dimensional functions. Recent advancements in computational tools to apply Markov Chain Monte Carlo (MCMC) methods are making Bayesian data analysis accessible for all statisticians. Two such computer tools are Win-BUGS and SASR 9.2's PROC MCMC. Bayesian methodology will be introduced through discussion of fourteen statistical examples with code and computer output to demonstrate the power of these computational tools in a wide variety of settings.

Educational Psychology
Ph.D.
Mathematics remains a subject many students fail to become competent in by the time they graduate from high school. Most students often require one on one, individualized tutoring to help them reach competence. That remains a challenge since most classrooms are understaffed and underfunded, frequently having only one teacher in a overpopulated classroom. One strategy that has been used to alleviate some of this over reliance on the teacher has been faded worked examples, or fading. Fading is the successive removal of the last steps in a series of problems until the student is solving problems completely on their own. The current study aimed to determine whether fading improves learning, and for whom. The goal was to compare fading with business as usual (control), worked examples with self-explanations, and fading with self-explanations. Specifically, I was interested in the following research questions: (1) Do the three experimental conditions differ in promoting posttest scores on surface area and volume? (2) Do the three experimental conditions differ in promoting conceptual knowledge and procedural knowledge of surface area and volume at posttest? and (3) When interaction terms are created between student profiles and conditions within regression analyses, which profiles explain significant variance in posttest scores? Repeated measures analysis of variance, principle axis factor analysis, and simple linear regressions were used to examine the differences between conditions at posttest, to create propensity scores, and to determine whether there were any interactions between propensity scores and conditions. Results indicated a significant effect of fading on posttest scores. A regression with propensity factors indicated that the fading conditions appeared to benefit low propensity students moreso than high propensity students. Findings are discussed in terms of educational implications and future research that can complement these findings to contribute to future research.
Temple University--Theses

Recent discoveries uncovered flaws in machine learning algorithms such as deep neural networks. Deep neural networks seem vulnerable to small amounts of non-random noise, created by exploiting the input to output mapping of the network. Applying this noise to an input image drastically decreases classication performance. Such image is referred to as an adversarial example. The purpose of this thesis is to examine how known regularization/robustness methods perform on adversarial examples. The robustness methods: dropout, low-pass filtering, denoising autoencoder, adversarial training and committees have been implemented, combined and tested. For the well-known benchmark, the MNIST (Mixed National Institute of Standards and Technology) dataset, the best combination of robustness methods has been found. Emerged from the results of the experiments, ensemble of models trained on adversarial examples is considered to be the best approach for MNIST. Harmfulness of the adversarial noise and some robustness experiments are demonstrated on CIFAR10 (The Canadian Institute for Advanced Research) dataset as well. Apart from robustness tests, the thesis describes experiments with human classification performance on noisy images and the comparison with performance of deep neural network.

In model-driven engineering, powerful query/view languages exist to compute result sets/views from underlying models. However, to use these languages effectively, one must understand the query/view language concepts as well as the underlying models and metamodels structures. Consequently, it is a challenge for domain experts to create queries/views due to the lack of knowledge about the computer-internal abstract representation of models and metamodels. To better support domain experts in the query/view creation, the goal of this paper is the presentation of a generic concept to specify queries/views on models without requiring deep knowledge on the realization of modeling languages. The proposed concept is agnostic to specific modeling languages and allows the query/view generation by-example with a simple mechanism for filtering model elements. Based on this generic concept, a generic query/view language is proposed that uses role-oriented modeling for its non-intrusive application for specific modeling languages. The proposed language is demonstrated based on the role-based single underlying model (RSUM) approach for AutomationML to create queries/views by-example, and subsequently, associated viewtypes to modify the result set or view.

Embodied cognition perspectives suggest that pointing and tracing with the index finger may support learning, with basic laboratory research indicating such gestures have considerable effects on information processing in working memory. The present thesis examined whether tracing worked examples could enhance learning through decreased intrinsic cognitive load. In Experiment 1, 56 Year 6 students (mean age = 11.20, SD = .44) were presented with either tracing or no-tracing instructions on parallel lines relationships. The tracing group solved more acquisition phase practice questions and made fewer test phase errors, but otherwise test results were limited by ceiling effects. 42 Year 5 students (mean age = 10.50, SD = .51) were recruited in Experiment 2 to better align the materials with students’ knowledge levels. The tracing group outperformed the non-tracing group at the test and reported lower levels of test difficulty, interpreted as lower levels of intrinsic cognitive load. Experiment 3 recruited 52 Year 6 and Year 7 students (mean age = 12.04, SD = .59) presented with materials on angle relationships of a triangle; the tracing effect was replicated on test scores and errors, but not test difficulty. Experiment 4 used the parallel lines materials to test hypothesized gradients across experimental conditions with 72 Year 5 students (mean age = 9.94, SD = .33), predicting the tracing on the paper group would outperform the tracing above the paper group, who in turn would outperform the non-tracing group. The hypothesized gradient was established across practice questions correctly answered, practice question errors, test questions correctly answered, test question time to solution, and test difficulty self-reports. The results establish that incorporating the haptic input into worked example-based instruction design enhances the worked example effect and that tracing worked examples is a natural, simple yet effective way to enhance novices’ mathematics learning.

Esta dissertação tem por objetivo analisar a atividade docente de Ferdinand de Saussure (1857-1913) relativa à Linguística Geral, em 1907, quando o então experiente professor assumiu, pela primeira vez, o ensino da disciplina (Curso I). Nossos pressupostos teóricos provêm da Historiografia Linguística, tal como praticada por Auroux (1994), Koerner (1996) e Swiggers (2004). Nossa problemática parte do reconhecimento de que uma das maneiras de perpetuar a presença de Saussure no cânone dos estudos linguísticos consiste em repetir metáforas e exemplos a ele creditados, como a célebre metáfora do jogo de xadrez para ilustrar o conceito de língua. A partir de estudos historiográficos das funções semiótica e argumentativa de definições e de exemplos de língua (Rey, 1995; Quijada Van den Berghe e Swiggers, 2009; Chevillard et al., 2007), repartimos o material de análise (Saussure, 1996[1907]) em três porções, a que chamamos domínios (Caussat, 1978): Fonologia, Fonética e Analogia, temas do Curso I, para os quais estabelecemos três objetivos principais de pesquisa [1 a 3], e um objetivo secundário [4]: [1] levantar e caracterizar definições nos três domínios; [2] levantar e caracterizar os exemplos de língua nos três domínios; [3] correlacionar o uso de definições e de exemplos de língua nos três domínios; [4] correlacionar os dados apurados em [3] com com o contexto imediato de emergência do Curso I, ou seja, com o ensino da disciplina Linguística Geral na Faculté de Lettres et Sciences Sociales, em Genebra, em 1907 (Vincent, 2013). Nossas análises, que abarcaram 60 (sessenta) definições, apontam para a preferência de se construir uma definição pelo termo (definitio nominis), e não pela coisa (definitio rei), quando se trata de apresentar definições prioritárias ao assunto. Quanto aos exemplos, foram apuradas mais de 2000 (duas mil) ocorrências, com saliente revezamento entre francês e alemão nos três domínios, acompanhados por latim e grego, em uma representação das línguas-objeto feita principalmente por compreensão, ou seja, pelo uso generalizável dos exemplos de língua, que ilustram o conteúdo temático de maneira não-exaustiva. A unidade majoritária dos exemplos é a palavra, e sua disposição espacial preferencial são séries paralelas, ora em contraste simultâneo, ora em relação de sucessão temporal, com poucos paradigmas e declinações. As perguntas retóricas desempenham papel didático frequente e flexível, desde o teste de hipóteses até a apresentação de definições. Observamos, ainda, a ocorrência de definições que apresentam diretamente exemplos de língua, em uma relação bastante forte de interdependência. Reunimos, ainda, mecanismos de organização dos conteúdos temáticos (numerais, grafemas alfabéticos) e de esquematização (tabelas, símbolos e outras rupturas com a linearidade da exposição oral e da anotação escrita). Por fim, observamos estruturas de negação em todos os mecanismos investigados: definições que apresentam aquilo que não é antes daquilo que é, exemplos de língua que não ilustram determinado assunto, perguntas retóricas com respostas frequentemente negativas e oposições entre conceitos, de modo que as relações de oposição, algumas antinômicas, parecem caracterizar de maneira geral a didática de Saussure antes mesmo da formulação explícita do conceito de signo linguístico, que ocorreu, como sabemos, no curso de 1910-1911.
This paper aims to examine the teaching activity of Ferdinand de Saussure (1857- 1913) as to General Linguistics, in 1907, when the experient professor first started a course on the subject. Our work assumes that one way of perpetrating the canon consists of repeating some metaphors and examples accredited to him, as it goes to the famous saussurian chess game metaphor which illustrates the concept of langue. Our theoretical background lies on the Historiography of Linguistics, as professed by Auroux (1994), Koerner (1996) and Swiggers (2004), and the same applies for our methodological assumptions (Rey, 1995). Having established three separate domains within the first course (Saussure, 1996[1907]), Phonology, Phonetics and Analogy, we have then settled three main goals [1 to 3] and one secondary objective [4]: [1] to describe the definitions given in the three domains; [2] to describe and assess the language examples provided by the three domains; [3] relate the uses of definitions to those of language examples; [4] to relate [3] to the immediate context of emergence of the first course, the institutional background of General Linguistics at the Faculté de Lettres et Sciences Sociales, (Geneva, 1907). In order to accomplish the first step [1], we have adopted the classification of definitio rei, definitio nominis, as far as the definition incidence is concerned, and essencial, formal and functional definition, as to its content (Quijada Van den Berghe e Swiggers, 2009). The language examples [2] were analysed through their démarcation, the way they appear in the text; the way the languages being described are represented in the text, which can assume two forms: either par extension, when the totality of the cases are listed, either par compréhension, when the examples are models from which one can generate new instances (Chevillard et al., 2007). Our corpus comprises 60 (sixty) definitions and over two thousand examples. In general terms, the definitio nominis is enhanced when it comes to define priority terms, such as phonology. As to the examples, four languages appear intensely in the domains: french, german, greek and latin. These languages entertain a kind of representation mainly in comprehension.These instances occur as words, in horizontal and parallel series, scarcely in paradigms or declensions. There are definitions which exemplify overtly and directly, with no more material to accomplish a definition, which shows the strict interdependence of these constructions, together with a different mechanism, the rethorical questions. We have gathered, as well, organizing mechanisms (items, topics) and schematization procedures (tables, symboles and strategies that lessen the linearity of the text). Eventually, we have spotted the wide-spread use of negative strategies, such as defining something by what it is not, or excluding elements from a definition, as well as providing language examples which uncorrectly illustrate what is being said, in order to anticipate possible mistakes. Such procedures seem to be a hallmark on Saussures teaching.

Large scale datasets collected using non-expert labelers are prone to labeling errors. Errors in the given labels or label noise affect the classifier performance, classifier complexity, class proportions, etc. It may be that a relatively small, but important class needs to have all its examples identified. Typical solutions to the label noise problem involve creating classifiers that are robust or tolerant to errors in the labels, or removing the suspected examples using machine learning algorithms. Finding the label noise examples through a manual review process is largely unexplored due to the cost and time factors involved. Nevertheless, we believe it is the only way to create a label noise free dataset. This dissertation proposes a solution exploiting the characteristics of the Support Vector Machine (SVM) classifier and the sparsity of its solution representation to identify uniform random label noise examples in a dataset. Application of this method is illustrated with problems involving two real-world large scale datasets. This dissertation also presents results for datasets that contain adversarial label noise. A simple extension of this method to a semi-supervised learning approach is also presented. The results show that most mislabels are quickly and effectively identified by the approaches developed in this dissertation.

As obras frequentemente consideradas como as mais relevantes para o estudo da filosofia kantiana são as três críticas que apresentam o rigor do pensamento filosófico e marcam uma mudança de perspectiva no pensamento moderno, apontando limites e novos horizontes para o conhecimento humano. Porém, concomitantemente ao desenvolvimento destas obras encontramos outras que também são importantes para a compreensão do projeto filosófico kantiano como um todo. Destaca-se no presente projeto de pesquisa a Antropologia de um ponto de vista pragmático e sua relevância para o estudo da formulação da noção moral a partir da observação das ações humanas e da possibilidade de identificação e formação do cidadão do mundo para Kant.
The works often regarded as the most relevant to the study of Kant\'s philosophy are the three critical presenting the rigor of philosophical thought and mark a change of perspective in modern thought, pointing limits and new horizons to human knowledge. However, concurrently with the development of these works find others that are also important for understanding the Kantian philosophical project as a whole. Stands out in this research project Anthropology from a pragmatic point of view and its relevance to the study of the formulation of the moral sense from the observation of human actions and the possibility of identifying and training the \"world citizen\" to Kant.

This report will deal with the process of creating an example based procedural distribution tool. This is accomplished within the Frostbite game engine editor, FrostEd. By using actual placements of objects in the editor as in-data, the tool provides the artist with an unmatched visual feel for calibrating its properties and settings. Note that this is a unique technique and was invented during the creation of this tool. The tool is based on a machine learning approach. It creates a feature vector from the example placements for each type of object. These vectors are then used to create statistical models which in turn are used to generate new object placements. The process of determining the position and rotation when generating an object is divided into two parts. A new concept called Feature Function (FF) is utilized to provide each element in the population with a probability to obtain a certain position and rotation.

Thesis (M.B.A.)--Naval Postgraduate School, December 2007.
Thesis Advisor(s): Abdel-Hamid, Tarek. "December 2007." Description based on title screen as viewed on January 24, 2008. Includes bibliographical references (p. 73). Also available in print.

This thesis considers possible criteria for the selection of example sentences for difficult or unknown words in reading texts for students of German as a Second Language (GSL). The examples are intended to be provided within the context of an Intelligent Computer-Aided Language Learning (ICALL) Vocabulary Learning System, where students can choose among several explanation options for difficult words. Some of these options (e.g. glosses) have received a good deal of attention in the ICALL/Second Language (L2) Acquisition literature; in contrast, literature on examples has been the near exclusive province of lexicographers. The selection of examples is explored from an educational, L2 teaching point of view: the thesis is intended as a first exploration of the question of what makes an example helpful to the L2 student from the perspective of L2 teachers. An important motivation for this work is that selecting examples from a dictionary or randomly from a corpus has several drawbacks: first, the number of available dictionary examples is limited; second, the examples fail to take into account the context in which the word was encountered; and third, the rationale and precise principles behind the selection of dictionary examples is usually less than clear. Central to this thesis is the hypothesis that a random selection of example sentences from a suitable corpus can be improved by a guided selection process that takes into account characteristics of helpful examples. This is investigated by an empirical study conducted with teachers of L2 German. The teacher data show that four dimensions are significant criteria amenable to analysis: (a) reduced syntactic complexity, (b) sentence similarity, provision of (c) significant co-occurrences and (d) semantically related words. Models based on these dimensions are developed using logistic regression analysis, and evaluated through two further empirical studies with teachers and students of L2 German. The results of the teacher evaluation are encouraging: for the teacher evaluation, they indicate that, for one of the models, the top-ranked selections perform on the same level as dictionary examples. In addition, the model provides a ranking of potential examples that roughly corresponds to that of experienced teachers of L2 German. The student evaluation confirms and notably improves on the teacher evaluation in that the best-performing model of the teacher evaluation significantly outperforms both random corpus selections and dictionary examples (when a penalty for missing entries is included).

As digital images become ubiquitous in many applications, the need for efficient and effective retrieval techniques is more demanding than ever. Query by Example (QBE) and Query by Concept (QBC) are among the most popular query models. The former model accepts example images as queries and searches for similar ones based on low-level features such as colors and textures. The latter model allows queries to be expressed in the form of high-level semantics or concept words, such as "boat" or "car," and finds images that match the specified concepts. Recent research has focused on the connections between these two models and attempts to close the semantic-gap between them. This research involves finding the best method that maps a set of low-level features into high-level concepts. Automatic annotation techniques are investigated in this dissertation to facilitate QBC. In this approach, sets of training images are used to discover the relationship between low-level features and predetermined high-level concepts. The best mapping with respect to the training sets is proposed and used to analyze images, annotating them with the matched concept words. One principal difference between QBE and QBC is that, while similarity matching in QBE must be done at the query time, QBC performs concept exploration off-line. This difference allows QBC techniques to shift the time-consuming task of determining similarity away from the query time, thus facilitating the additional processing time required for increasingly accurate matching. Consequently, QBC's primary design objective is to achieve accurate annotation within a reasonable processing time. This objective is the guiding principle in the design of the following proposed methods which facilitate image annotation: 1.A novel dynamic similarity function. This technique allows users to query with multiple examples: relevant, irrelevant or neutral. It uses the range distance in each group to automatically determine weights in the distance function. Among the advantages of this technique are higher precision and recall rates with fast matching time. 2.Object recognition based on skeletal graphs. The topologies of objects' skeletal graphs are captured and compared at the node level. Such graph representation allows preservation of the skeletal graph's coherence without sacrificing the flexibility of matching similar portions of graphs across different levels. The technique is robust to translation, scaling, and rotation invariants at object level. This technique achieves high precision and recall rates with reasonable matching time and storage space. 3.ASIA (Automatic Sampling-based Image Annotation) is a technique based on a new sampling-based matching framework allowing users to identify their area of interest. ASIA eliminates noise, or irrelevant areas of the image. ASIA is robust to translation, scaling, and rotation invariants at the object level. This technique also achieves high precision and recall rates. While the above techniques may not be the fastest when contrasted with some other recent QBE techniques, they very effectively perform image annotation. The results of applying these processes are accurately annotated database images to which QBC may then be applied. The results of extensive experiments are presented to substantiate the performance advantages of the proposed techniques and allow them to be compared with other recent high-performance techniques. Additionally, a discussion on merging the proposed techniques into a highly effective annotation system is also detailed.
Ph.D.
School of Computer Science
Engineering and Computer Science
Computer Science

L'exemple és un component necessari de tot diccionari de llengua, i especialment d'un bilingüe destinat a aprendre i usar una llengua estrangera. En aquest treball, després d'establir les bases teòriques i la terminologia usada (la lexicografia bilingüe, la concepció de l'diccionari com a discurs metalingüístic de el lèxic i com a text condensat, els conceptes fonamentals de la metalexicografía referits a la microestructura), es determina la naturalesa i l'estatus teòric de l'exemple lexicogràfic, i més en concret, de l'bilingüe; s'ofereix una definició exhaustiva de l'exemple i s'especifiquen els trets distintius de l'exemple bilingüe que es desenvolupen a continuació. Es determinen els tipus d'exemples lexicogràfics, prestant una major atenció als de diccionaris bilingües. S'estudien les funcions que exerceixen els exemples (que són bàsicament les d'il·lustrar la informació oferta en l'article i aportar una informació complementària), el seu valor de model a partir de qual l'usuari pot formar sintagmes o oracions i la seva major o menor vinculació a l' sistema lingüístic (langue) o realitzacions pròpies de la parla (parole). L'exemple bilingüe no només il·lustra la paraula lematitzada sinó l'ús de l'equivalent, per això es distingeixen, de la categoria lexicogràfica de l'exemple, les unitats lèxiques fixes que en molts diccionaris bilingües es presenten tipogràficament com a exemples, però que en realitat no exemplifiquen, sinó que poden ser exemplificades: el infralema (subentrada), categoria el contingut es defineix en termes precisos (com infralema es presenten no només locucions, sinó també combinacions estables o peces lèxiques i pragmatemas que tenen un significat precís i, en la majoria dels casos , un equivalent canònic en una altra llengua). També s'analitzen els exemples la traducció no inclou l'equivalent presentat o que il·lustren un ús de la paraula lematitzada de la qual no s'ha donat tan sols equivalent. Atès que l'exemple il·lustra l'ús de la paraula lematitzada des de l'aspecte gramatical i lèxic, són objecte d'estudi exhaustiu les informacions gramaticals i lèxiques que aporta l'exemple. A més de les explicitacions morfològiques que pot revelar l'exemple, l'exemple mostra les implicacions sintàctiques de la paraula i aquí s'analitza quan l'exemple il·lustra una informació ja donada a l'article i quan aquesta ha de ser explicitada per l'usuari. El comportament combinatori de la paraula lematitzada és la informació més freqüent que aporta l'exemple, pel que es presta a aquest aspecte una atenció especial, començant pel mateix concepte de col·locació, que en l'àmbit de la lexicografia és força indeterminat. S'assenyala, però, que la funció de mostrar la combinatòria lèxica no és exclusiva de l'exemple, ja que la comparteix amb la indicació de cotext. També s'examina un tipus d'exemple, l'anomenat hipercondensado, que comparteix trets de l'exemple i de l'indicador de cotext. També s'analitza la informació pragmàtica que pot aportar l'exemple. Tots els aspectes relatius a l'exemple estan degudament il·lustrats amb mostres de diccionaris. No obstant això s'aplica tot l'aparell teòric exposat a l'anàlisi d'un corpus format per 3.835 segments presentats com a exemples de la lletra L dels diccionaris lituà-català / italià / francès. Es fa una primera distinció entre aquelles seqüències que corresponen al que en la primera part hem determinat com a exemple lexicogràfic i les que corresponen a la categoria de infralema. Després es divideixen els exemples en tres grans grups (oracions finites, oracions infinitivas i sintagmes no oracionals) i dins de cada grup es divideixen al seu torn en subgrups i s'analitzen els següents aspectes: quin tipus d'informació (gramatical, lèxica, pragmàtica) aporta o pot aportar cada tipus d'exemple, si l'exemple il·lustra o complementa la informació.
El ejemplo es un componente necesario de todo diccionario de lengua, y especialmente de un bilingüe destinado a aprender y usar una lengua extranjera. En este trabajo, después de establecer las bases teóricas y la terminología usada (la lexicografía bilingüe, la concepción del diccionario como discurso metalingüístico del léxico y como texto condensado, los conceptos fundamentales de la metalexicografía referidos a la microestructura), se determina la naturaleza y el estatus teórico del ejemplo lexicográfico, y más en concreto, del bilingüe; se ofrece una definición exhaustiva del ejemplo y se especifican los rasgos distintivos del ejemplo bilingüe que se desarrollan a continuación. Se determinan los tipos de ejemplos lexicográficos, prestando una mayor atención a los de diccionarios bilingües. Se estudian las funciones que desempeñan los ejemplos (que son básicamente las de ilustrar la información ofrecida en el artículo y aportar una información complementaria), su valor de modelo a partir del cual el usuario puede formar sintagmas u oraciones y su mayor o menor vinculación al sistema lingüístico (langue) o a realizaciones propias del habla (parole). El ejemplo bilingüe no solo ilustra la palabra lematizada sino el uso del equivalente, por eso se distinguen, de la categoría lexicográfica del ejemplo, las unidades léxicas fijas que en muchos diccionarios bilingües se presentan tipográficamente como ejemplos, pero que en realidad no ejemplifican, sino que pueden ser ejemplificadas: el infralema (subentrada), categoría cuyo contenido se define en términos precisos (como infralema se presentan no solo locuciones, sino también combinaciones estables o piezas léxicas y pragmatemas que tienen un significado preciso y, en la mayoría de los casos, un equivalente canónico en otra lengua). También se analizan los ejemplos cuya traducción no incluye el equivalente presentado o que ilustran un uso de la palabra lematizada de la que no se ha dado siquiera equivalente. Dado que el ejemplo ilustra el uso de la palabra lematizada desde el aspecto gramatical y léxico, son objeto de estudio exhaustivo las informaciones gramaticales y léxicas que aporta el ejemplo. Además de las explicitaciones morfológicas que puede revelar el ejemplo, el ejemplo muestra las implicaciones sintácticas de la palabra y aquí se analiza cuándo el ejemplo ilustra una información ya dada en el artículo y cuándo esta ha de ser explicitada por el usuario. El comportamiento combinatorio de la palabra lematizada es la información más frecuente que aporta el ejemplo, por lo que se presta a este aspecto una atención especial, empezando por el mismo concepto de colocación, que en el ámbito de la lexicografía es bastante indeterminado. Se señala no obstante que la función de mostrar la combinatoria léxica no es exclusiva del ejemplo, ya que la comparte con la indicación de cotexto. También se examina un tipo de ejemplo, el llamado hipercondensado, que comparte rasgos del ejemplo y del indicador de cotexto. También se analiza la información pragmática que puede aportar el ejemplo. Todos los aspectos relativos al ejemplo están debidamente ilustrados con muestras de diccionarios. No obstante se aplica todo el aparato teórico expuesto al análisis de un corpus formado por 3.835 segmentos presentados como ejemplos de la letra L de los diccionarios lituano-español/italiano/francés. Se hace una primera distinción entre aquellas secuencias que corresponden a lo que en la primera parte hemos determinado como ejemplo lexicográfico y las que corresponden a la categoría de infralema. Después se dividen los ejemplos en tres grandes grupos (oraciones finitas, oraciones infinitivas y sintagmas no oracionales) y dentro de cada grupo se dividen a su vez en subgrupos y se analizan los siguientes aspectos: qué tipo de información (gramatical, léxica, pragmática) aporta o puede aportar cada tipo de ejemplo, si el ejemplo ilustra o complementa la información.
The example is an essential component of a dictionary of any language, and especially of a bilingual one intended to learn and use a foreign language. In this thesis, after stablishing the theoretical basis and the terminology used (bilingual lexicography, the conception of the dictionary as a metalinguistic discourse of the lexicon and as a condensed text, the fundamental concepts of metalexicography referred to the microstructure), the nature and the theoretical status of lexicographic example, and more specifically, the bilingual one, is determined; a comprehensive definition of the example is provided and the distinctive features of the bilingual example − developed below − are specified. The types of lexicographic examples are specified, paying a greater attention to those of bilingual dictionaries. The functions of the examples (which are basically to illustrate the information offered in the article and provide complementary information), their value as a model from which the user can form phrases or sentences, and their greater or lesser links to the linguistic system (langue) or realizations of speech (parole) are studied. The bilingual example not only illustrates the lemmatized word but may also illustrate the use of the equivalent, for thar reason, the fixed lexical units, which in many bilingual dictionaries are typographically presented as examples, but in fact do not exemplify and may be exemplified, are distinguished from the category of the dictionary example: the infralemma (subentry), a category whose content is defined in precise terms (not only idioms but also stable lexical combinations or lexical pieces and pragmathemes that have a precise meaning and, in most cases, a canonical equivalent in another language, are presented as infralemmas). Another important topic is also analyzed - examples whose translation does not contain the given equivalent, or which illustrate the use of the lemmatized word, even if no equivalent is given. Since the example illustrates the use of the lemmatized word in the grammatical and lexical aspect, the grammatical and lexical information provided by the example is the object of an exhaustive study. In addition to the morphological explanations that the example may reveal, it shows the syntactic implications of the word and, in this thesis, it is analyzed when the example illustrates information already given in the article and when it has to be made explicit by the user. The combinatorial behavior of the lemmatized word is the most frequent information in the example, and hence special attention is paid to this aspect, starting with the concept of colocation, which is quite indeterminate in the field of lexicography. It should be noted that the lexical-combinatory information is not exclusive to the example, since the cotext indication also conveys this information. A type of example, the so-called hypercondensed one, which shares features with the example and the cotext indication is examined as well. The pragmatic information that the example may provide is also discussed. All aspects related to the example are thoroughly illustrated with samples of dictionaries. However, the entire proposed theoretical apparatus is applied to the analysis of a corpus comprised of 3 835 segments presented as examples of the letter L of the Lithuanian-Spanish/Italian/French dictionaries. A first distinction is made between those sequences that correspond to what has previously been determined as a lexicographic example and those that correspond to the category of infralemma. The examples are then divided into three large groups (finite sentences, infinitive clauses and non-sentence phrases) and, within each group, they are in turn divided into subgroups and then the following aspects are analyzed: what type of information (grammatical, lexical, pragmatic) each type of example provides or may provide, whether the example illustrates or complements the information already given.
Universitat Autònoma de Barcelona. Programa de Doctorat en Llengües i Cultures Romàniques

This diploma thesis deals with the business property and its meaning in the Czech and Slovak legislations. First of all, a historical background of the business property definition is presented. This term is then described within scope of particular laws where it is defined in slightly different ways. There are mentioned also differences between the Czech and Slovak legislations. Selected judicial cases are used to demonstrate practical applications of the business property term. There is mentioned also the problem of intangible assets protection with no judicial precedent so far. Conceptual differences of the business property meaning are highlighted in a model example. Results, interpretation and comparison of questionnaire investigations both from the Czech Republic and Slovakia are an important part of this work.

Program synthesis is the task of automatically finding a program in an underlying Domain-Specific Language (DSL) that satisfies a specification provided by the user. In recent years the topics of program synthesis and in particular programming-by-example (PBE), a branch of program synthesis that uses input-output examples as specification, have gained significant traction, also thanks to the use of PBE tools in commercial applications, such as Excel'sFlashFill. Program synthesis is often considered as a difficult search problem, and one of the main focus of synthesis algorithms is to reduce and efficiently explore the search space. In addition to these techniques some synthesizers parallelize the search to improve performance, but the benefits of this optimization is still severely limited. In this thesis I explore the possibility of using distributed systems to further improve the performance of synthesis algorithms by increasing the number of parallel searches. I introduce an algorithm for PBE that aims to efficiently parallelize the search over a distributed system and implement it into a framework, then analyze the performance of the framework and which characteristics influence the effectiveness of distributed synthesis using instantiations for the string transformation and SQL domains. The evaluation of these instantiations on existing benchmarks shows that the developed algorithm allows performance comparable to that of current state-of-the-art techniques when used in a serial manner, and can also obtain a speedup of over 3x by using 5 parallel searches over a small distributed system.

The aim of the thesis is to study the pseudonymous letters of Aeschines, all of which purport to give an account of his sojourn in exile. There is a strong consensus among scholars that all the letters are forgeries, and their date of composition tends to be located in the first few centuries CE on linguistic grounds. Embracing a variety of literary forms, these letters were probably composed by multiple hands and may for convenience be divided into three categories: Letters 2, 3, 7, 11, 12 imitate the ‘Demosthenic’ letters in a manner similar to the Hellenistic (and beyond) historical declamations and progymnasmata; Letters 1, 5, 6, 8, 9, 10 come to us with features reminiscent of what German scholars would call Briefromane, or ‘epistolary novels’, and are normally deemed typical of the so-called Second Sophistic; and Letter 4 is a showpiece assuming the form of a Pindaric exegesis. The thesis consists of two parts. The first gives an extensive account of the letters, including their background, history of scholarship, and basic features, to seek to present the ‘forger’ and the text in their proper historical and cultural contexts. The second part, which constitutes the basis for the reflections developed in the first, provides a detailed commentary in thematic sequence. It begins with the ‘Demosthenic’ counterparts (Epp. 2, 3, 7, 11, 12), and stylistic comparisons are made throughout. The analysis of the fictional letters (Epp. 1, 5, 6, 8, 9) pays particular attention to their consistency of narrative and engagement with other literary genres. The commentary on Letter 4 foregrounds the Pindaric elements and completes the thesis. Letter 10 is discussed at sporadic points: it is a later attachment to the corpus and the erotic content is inconsistent with the ‘original’ forgeries. The overall focus of the thesis is on two overlapping aspects of Aeschines’ early reception in antiquity – as ‘the other orator’ beside Demosthenes and as inspiration for later rhetorical education. Existing studies, however, are more concerned with textual criticism and linguistic analysis and have left the letters almost unproductive in these respects: so Drerup (1904), Schwegler (1914), and, most recently, García Ruiz and Hernández Muñoz (2012). In his classic work Goldstein (1968) took the parallel passages in the pseudonymous letters as evidence for authenticating Demosthenes’ letters, and scholars are now able to take advantage of a more reliable reference when studying Ps.-Aeschines. Holzberg (1994), on the other hand, established a set of generic criteria for the Briefromane and has substantially changed the way we read Ps.-Aeschines: it is now possible to appreciate the literary value of the letters without scrutinising their authenticity. Yet both these studies tell us only half the story: while Goldstein left more remarks on the imitative counterparts of Demosthenes’ letters, Holzberg focused on the way the letters reflect the epistolary narrative. Following Rohde (1876/1960), moreover, it seems common sense to characterise the pseudo-historical tale as seen through the letters as a product of the ‘Second Sophistic’, though discoveries of new papyri, e.g. the Ninus romance c. first century BCE, undermined this assumption. My study is built on these investigations in an attempt to form the most extended analysis. The study of the ‘Demosthenic’ counterparts will contribute to a better understanding of Ps.-Aeschines’ intertextual engagement with Demosthenes and his successors, e.g. Ps.-Leosthenes (FGrH 105 F 6 = MP3 2496). It shows that Ps.-Aeschines owes a great deal to the culture of rhetoric and highlights his significance in the Nachleben of Attic oratory. As for the other letters, this thesis argues that they deserve some space in our accounts of the history of exilic, periegetic, and epinician literatures for contextualising a wide range of preexisting literary forms such as the Homeric Odyssey (Ep. 1) and Pindar’s victory odes (Ep. 4). As contingent by-products of the ‘Demosthenic’ counterparts, however, they seem to allow no confident judgement about generic consciousness, esp. the very notion of ‘novel’, and need to be approached as antedating the Imperial exponents. Contrary to the communis opinio, therefore, I attempt to move the date of composition forward to the late Hellenistic period, in which there was already ample encouragement for a sophist, as well as for his students, to write pseudonymous letters. The ‘traitors’ blacklist’ (Ep. 12.8–9) and the term for the Rhodian family of Diagoreans (Ep. 4.4) entertain this possibility inasmuch as both show marked affinities with the Hellenistic sources. Last but not least, the two coexisting, radically opposed interpretations of one’s civic orientation in exile will help us tackle the stability and change in the political cultures of the post-Classical era. My conclusion is that these letters hold a unique position as very early – and very illuminating – examples of how different literary, political trends were interwoven to make, and to remould, a Classic. It is hoped that this study may have done something to reappraise Ps.-Aeschines, who is, in all likelihood, a pre-sophisticated forerunner at a crossroads in the history of Greek literature.

The use of new tools for mathematics at school wins increasingly Importance. It follows from this that they are consquences as well as on aims and contents of mathematics at school us like on methods in the lessons. It is not unusual, that students and parents and also university professors are to be feared, that the calculation competence is decreasing with the use of CAS. In the lesson should be showed a possible way to developing such competences in the beginning phase of the learning process in Algebra. The examples refer to a level of school mathematics for students in the middle school age. The methods tell apart phases with and without CAS and shows a didactic principles of mathematics lessons in the case of use of CAS handheld technology.

The example based processing problem can be expressed as: "Given an example of an image or video before and after processing, apply a similar processing to a new image or video". Our thesis is that there are some problems where a single general algorithm can be used to create varieties of outputs, solely by presenting examples of what is desired to the algorithm. This is valuable if the algorithm to produce the output is non-obvious, e.g. an algorithm to emulate an example painting's style. We limit our investigations to example based processing of images, video, and 3D models as these data types are easy to acquire and experiment with. We represent this problem first as a texture synthesis influenced sampling problem, where the idea is to form feature vectors representative of the data and then sample them coherently to synthesize a plausible output for the new image or video. Grounding the problem in this manner is useful as both problems involve learning the structure of training data under some assumptions to sample it properly. We then reduce the problem to a labeling problem to perform example based processing in a more generalized and principled manner than earlier techniques. This allows us to perform a different estimation of what the output should be by approximating the optimal (and possibly not known) solution through a different approach.

Estudamos exemplos de métricas Finsler simétricas e não-simétricas em S^n, CP^n e HP^n com uma quantidade finita de geodésicas fechadas ou com uma quantidade pequena de geodésicas fechadas \"curtas\". São os chamados exemplos de Katok. Usamos como referência o artigo \"Geometry of the Katok examples\" de Wolfgang Ziller. Verificamos que existem métricas Finsler cujo número de geodésicas fechadas é 2n (no caso de S^ e S^), n(n+1) (no caso de CP^n) e 2n(n+1) (no caso de CP^n). Tais exemplos são construídos numa vizinhança qualquer da métrica Riemanniana canônica dessas variedades.
We study examples of symmetric and non-symmetric Finsler metrics on S^n, CP^n and HP^n with a finite number of closed geodesics or with a small number of \"short\" closed geodesics. These are the well known Katok\'s examples. We use Ziller\'s article Geometry of the Katok examples. We exhibit Finsler metrics whose number of closed geodesics is 2n (in the case of S^ and S^), n(n+1) (in the case of CP^n) and 2n(n+1) (in the case of HP^n). Such examples are found in any neighborhood of the canonical Riemannian metric on these manifolds.

In the thesis, we study improvements in the precision of inefficiency estimation by considering the dependence information among the error terms in models of production. We propose that the production systems that account for technical and allocative inefficiencies offer a natural way to model dependence using vine copulas. We construct such vine copulas using a recently proposed family of bivariate copulas (APS-2 copula) that permit dependence between the magnitude (but not the sign) of the allocative inefficiency and the magnitude of the technical inefficiency, and a Gaussian copula. We show how to estimate such models and argue that they better reflect dependencies that arise in practice. Following that we study around 200 features of personal loan customers, and apply machine learning models to see how these characteristics can help with risk measurement. The cost for different misclassification errors are different for personal loan providers. Furthermore, the misclassification costs for each example are different. Therefore, we propose an example-dependent cost-sensitive gradient boosting model and apply the method to a personal loan dataset.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-07-19T17:46:58Z No. of bitstreams: 1 rafaeladominguescosta.pdf: 9722437 bytes, checksum: b36e3a69cc20942105f6d3e9d3df9cd8 (MD5)
Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-08-09T11:53:07Z (GMT) No. of bitstreams: 1 rafaeladominguescosta.pdf: 9722437 bytes, checksum: b36e3a69cc20942105f6d3e9d3df9cd8 (MD5)
Made available in DSpace on 2017-08-09T11:53:07Z (GMT). No. of bitstreams: 1 rafaeladominguescosta.pdf: 9722437 bytes, checksum: b36e3a69cc20942105f6d3e9d3df9cd8 (MD5) Previous issue date: 2008-05-08
CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
Os estudos em torno de por exemplo, inicialmente, revelaram que esse objeto caracterizava-se pela função de conector apositivo, interligando unidade base e unidade apositiva. Essa função despertou o nosso interesse para a pesquisa e nos levou a averiguar, sincronicamente, as possíveis funções e comportamentos de por exemplo. Assim, durante as investigações dos dados, encontramos ocorrências de por exemplo apositivo e por exemplo não-apositivo, ambos apresentando uma característica extremamente peculiar: a mobilidade. Sendo assim, quando apositivo, por exemplo ocorre em a) posição inicial, encabeçando unidade apositiva; b) em posição medial, após sintagma (nominal, preposicionado ou adverbial) introdutor de aposição, incidindo em fronteiras de constituintes "...sujeito" e sujeito-verbo. E, ainda, e) nas fronteiras de constituintes (verbo-complemento e fronteira final) que, além de realizar a função focalizadora, apresenta as características simbólicas da aposição, como a relação semântica geral-específico e a correferencialidade parcial. Já em ocorrências não-apositivas, por exemplo é encontrado em fronteiras de constituintes ("...- sujeito", sujeito-verbo, verbo-complemento e, ainda, em fronteira final), desempenhando apenas função focalizadora, isto é, ressaltando anaforicamente ou cataforicamente (dependendo da posição do sintagma preposicionado) uma informação nova. Motivados pela multifuncionalidade desse fenômeno, buscamos também examinar, em âmbito diacrônico, sua origem e trajetória, a fim de legitimar tal diversidade funcional sincrônica apresentada nesse trabalho. Para que alcançássemos tal objetivo de pesquisa, baseamos-nos, principalmente, nos parâmetros de estratificação, divergência, especialização, persistência e decategorização, postulados por Hopper (1991), no pressuposto de unidirecionalidade de Heine et alli (1991), em que itens lexicais básicos ao se gramaticalizarem tornam-se abstratos. E, ainda, na gramaticalização como processo metafórico orientada por estágios, sugerido por Heine (1993). E, foi com base nesses pressupostos, que se procurou apresentar os aspectos da mudança desse fenômeno e indicar a hierarquização de seus usos. Para a realização de tal investigação, utilizaram-se corpora sincrônicos relacionados à língua falada dos projetos: "Fala Mineira", "Corpus de Conceição de Ibitipoca" e dados do "Projeto PROCON". Selecionamos também, como amostras de língua escrita, as seções de entrevista da Revista Cláudia, as páginas amarelas da Revista Veja/ Roberto Pompeu e artigos da Revista Veredas. Quanto aos corpora diacrônicos, contamos com o Corpus Diacrônico do Português, no entanto, tivemos acesso a apenas os séculos XIII, XIV, XV, XVII E XX. Examinamos, também, o Corpus Informatizado do Português Medieval, séculos XIV, XV, XVI, XVII, XVIII e XX; e, ainda, elegemos alguns textos avulsos dos séculos XVI, XVIII E XIX. Tais dados foram submetidos ao programa estatístico GOLDVARB (2001), cujos resultados foram extremamente significantes para comprovar as diversas funções atribuídas ao por exemplo.
Studies around "for example", initially, showed that this object characterized by the function of appositive connector, linking base unit and appositive unit. This function awakened our interest to the research and led us to investigate, synchronically, the possible functions and behaviors of "for example". Thus, during the investigations of our data, find occurrences of such appositive and such non-appositive, both presenting an extremely peculiar characteristic: mobility. So when appositive, "for example" occurs in a) starting position, introducing appositive unit b) in the medial position, after words (nominal, propositioned or adverbial) introducer of affixing, and c) the border of constituents (subject - verb, verb-completion and final frontier), which, in addition to perform the focus function, has the characteristics of affixing symbolic, as the relationship-specific and general semantics partial correfereciality. In instances non-appositives, "for example" is found in borders of constituents ("...- subject", subject-verb, verb-complement, supplement, complement, and in the final frontier), playing only :focus function, ie emphasizing anaphorically or cataphorically (depending's position propositioned words) a new information. Driven by the multifunctionality of this phenomenon, we sought also examine, in part diachronic, its origins and trajectory, in order to legitimize such functional diversity synchronous presented in this work. To reach the objective of this research, base ourselves, mainly on the principles of stratification, divergence, expertise, persistence and decategorizacao, postulates by Hopper (1991), on the assumption unidirecionalidade of Heine (1991), in which the basic lexical items if gramaticalizarem become abstract. And yet, in the grammaticalization as metaphoric process guided by stages, suggested by Heine (1993). And, it was based on these assumptions, which sought present aspects of the change of this phenomenon and indicate the hierarchy of its uses. For the achievement of such research, used up corpora synchronous related to the language of the projects: "Speech Mining", "Corpus of Conception of Ibitipoca" and data of the "Project PROCON." Also selected as samples of written language, the sections of the Magazine interview Claudia, the yellow pages of Revista Veja / Roberto Pompeu and of the journal Veredas. For diachronic corporate, count on the Corpus Diacronico do Portugues, however, have access to only the centuries XIII, XIV, XV, XVII and XX. Looked, too, the Corpus Informatizado do Portugues Medieval, centuries XIV, XV, XVI, XVII, XVIII and XX; and elected a brief avulses texts to the centuries XVI, XVII and XIX. These data were submitted to the statistical programme GOLDVARB (2001), whose results were very significant to demonstrate the various tasks assigned to eg.

The thesis focuses on the discriminatory treatments in the Czech tax law. The aim is to analyze the Czech Income Tax Act and the Value Added Tax Act to identify possible tax discriminations. The thesis studies infringement procedures leaded by the European Commission and it also studies cases of the European Commission against the Czech Republic on this matter and cases that can have impact on the Czech Republic because non discrimination is one of the primary principles of EU law. The identified discriminations should be removed in such a way that the Czech tax legislation does not discriminate against any group of taxpayers

We propose a running example for heterogeneous approach based on new type of fuzzification that diversifies fuzziness of every object, fuzziness of every attribute and fuzziness of every table value in a formal context. Moreover we suggest another working examples on heterogeneous environment and provide additional utilization and illustration of this new model that allows to use Formal Concept Analysis also for heterogenenous data. An interpretation of heterogeneous formal concepts and the resulting concept lattice is included.

本文データは平成22年度国立国会図書館の学位論文(博士)のデジタル化実施により作成された画像ファイルを基にpdf変換したものである
Kyoto University (京都大学)
0048
新制・論文博士
博士(情報学)
乙第10698号
論情博第16号
新制||情||9(附属図書館)
UT51-2001-F496
(主査)教授 池田 克夫, 教授 佐藤 雅彦, 教授 美濃 導彦
学位規則第4条第2項該当

We propose a running example for heterogeneous approach based on new type of fuzzification that diversifies fuzziness of every object, fuzziness of every attribute and fuzziness of every table value in a formal context. Moreover we suggest another working examples on heterogeneous environment and provide additional utilization and illustration of this new model that allows to use Formal Concept Analysis also for heterogenenous data. An interpretation of heterogeneous formal concepts and the resulting concept lattice is included.

A perturbation analysis shows that if a numerically stable procedure is used to compute the matrix sign function, then it is competitive with conventional methods for computing invariant subspaces. Stability analysis of the Newton iteration improves an earlier result of Byers and confirms that ill-conditioned iterates may cause numerical instability. Numerical examples demonstrate the theoretical results.