Accounting guidance often contains examples which provide practitioners with a description of a hypothetical transaction and its appropriate accounting treatment. Despite this potential to influence accounting judgments, our understanding of how practitioners use such examples when making these judgments is limited. Relying on psychology theory, I propose that practitioners must first assess the level of similarity between the transaction and the example. I predict that when doing this, practitioners unknowingly use a biased cognitive process where they overweight shared aspects between the transactions. Using an experiment, I confirm this prediction and show that this bias causes practitioners to systematically assess similarity between a transaction and example as too high. Results also show that this causes practitioners to consistently overestimate the likelihood that their transaction also qualifies for the same treatment as any example they are given. My study provides insights on how and why examples can systematically affect accounting judgments and has implications for both standard setters and practitioners. / text

Examples are a powerful tool for teaching both humans and computers.In order to learn from examples, however, a student must first extractthe examples from its stream of perception. Snapshot learning is ageneral approach to this problem, in which relevant samples ofperception are used as examples. Learning from these examples can inturn improve the judgement of the snapshot mechanism, improving thequality of future examples. One way to implement snapshot learning isthe Top-Cliff heuristic, which identifies relevant samples using ageneralized notion of peaks. I apply snapshot learning with theTop-Cliff heuristic to solve a distributed learning problem and showthat the resulting system learns rapidly and robustly, and canhallucinate useful examples in a perceptual stream from a teacherlesssystem.

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.

While exploring unfamiliar concepts and striving to grasp higher level mathematics, secondary and postsecondary mathematics students often ask, “When will we ever use this?” Although this question typically stems from students’ frustration, skepticism, and confusion, the question has great potential for teachable moments. Mathematics has countless applications in people’s daily lives, but the common person often fails to recognize this; those who realize the worldly importance of applicable mathematics often cannot provide specific examples nor understand the rigorous mathematics involved. It is important for mathematics teachers to have a conceptual understanding of the subject, and to be able to provide specific examples of applicable mathematics to students. Although the limit of applicable mathematics examples is infinite, a few cases are explored in this report. / text

<p><p>The use of examples is known to be important in learning, they should be “exemplary” and function as role-models.</p><p>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.</p><p>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.</p><p>The He[d]uristics are discussed and exemplified and also evaluated versus the derived set of characteristics and known common problems experienced by novices.</p></p>

We present a controlled experiment for the empirical evaluation of Example-Driven Modeling (EDM), an approach that systematically uses examples for model comprehension and domain knowledge transfer. We conducted the experiment with 26 graduate and undergraduate students from electrical and computer engineering (ECE), computer science (CS), and software engineering (SE) programs at the University of Waterloo. The experiment involves a domain model, with a UML class diagram representing the domain abstractions and UML object diagrams representing examples of using these abstractions. The goal is to provide empirical evidence of the effects of suitable examples on model comprehension, compared to having model abstractions only, by having the participants perform model comprehension tasks. Our results show that EDM is superior to having model abstractions only, with an improvement of (+39%) for diagram completeness, (+30%) for study questions completeness, (+71%) for efficiency, and a reduction of (-80%) for the number of mistakes. We provide qualitative results showing that participants receiving model abstractions augmented with examples experienced lower perceived difficulty in performing the comprehension tasks, higher perceived confidence in their tasks' solutions, and asked fewer clarifying domain questions (a reduction of 90%). We also present participants' feedback regarding the usefulness of the provided examples, the number of examples, the types of examples, and the use of partial examples.

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.

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.

Educators in the current era face more pressure to meet learners' growing digital age learning needs, which may require fostering more vital computational thinking skills. To ensure the desired learning outcomes are attained, it is critical to know how to provide the appropriate type of guidance and assistance. The findings of this research may be significant to computer science instructors and instructional designers interested in fostering computational thinking skills and improving programming skills by designing effective worked examples. Following the integrative review methodology, the study examined the current literature on worked examples in a programming setting to determine the compelling designs of worked examples. In addition, this study examined the most employed instructional design principles in developing effective worked examples and explored factors and circumstances that may have impacted the effectiveness of those designs. This study's findings indicated several successful designs of worked examples to promote computational thinking skills in programming problems / Doctor of Philosophy / Educators focus on fulfilling learners' expanding digital age learning requirements, which may require developing more critical computational thinking skills. It is vital to understand how to give appropriate guidance and support to achieve the intended learning results in programming courses. The outcomes of this study may be helpful to computer science educators and instructional designers who aim to support learners in gaining more advanced computational thinking skills. The study used the integrative review approach to investigate the current literature on worked examples in a programming context to discover the compelling designs of worked examples. The study provides information about the factors that may affect the design in addition to discuss several instructional design principles in regard to worked examples. The outcomes of this study showed numerous successful designs of worked examples that are helping in enhancing computational thinking skills in programming tasks.

<p>Computer vision applications such as image classification and object detection often suffer from adversarial examples. For example, adding a small amount of noise to input images can trick the model into misclassification. Over the years, many defense mechanisms have been proposed, and different researchers have made seemingly contradictory claims on their effectiveness. This dissertation first presents an analysis of possible adversarial models and proposes an evaluation framework for comparing different more powerful and realistic adversary strategies. Then, this dissertation proposes two randomness-based defense mechanisms Random Spiking (RS) and MoNet to improve the robustness of image classifiers. Random Spiking generalizes dropout and introduces random noises in the training process in a controlled manner. MoNet uses the combination of secret randomness and Floyd-Steinberg dithering. Specifically, input images are first processed using Floyd-Steinberg dithering to reduce their color depth, and then the pixels are encrypted using the AES block cipher under a secret, random key. Evaluations under our proposed framework suggest RS and MoNet deliver better protection against adversarial examples than many existing schemes. Notably, MoNet significantly improves the resilience against transferability of adversarial examples, at the cost of a small drop in prediction accuracy. Furthermore, we extend the usage of MoNet to the object detection network and use it to align with model ensemble strategies (Affirmative and WBF (weighted fusion boxes)) and Test Time Augmentation (TTA). We call such a strategy 3MIX. Evaluations found that 3Mix can significantly improve the mean average precision (mAP) on both benign inputs and adversarial examples. In addition, 3Mix is a lightweight approach to migrate the adversarial examples without training new models.</p>