Doctor of Philosophy / Department of Educational Leadership / Sarah Jane Fishback / Today, educators try to maximize eLearning effectiveness by using various methodologies to engage the learner and increase learning and recall. If educators want to be able to continue this process, new eLearning methodologies must be investigated. The aim of this study was to determine if the method in which lesson content is present can affect learning. To this end, the research question was as follows: What difference, if any, does the random placement of small paragraphs, chunks, of text make in recall and learning in eLearning lessons? The research question was answered through an experiment that entails assigning an eLearning lesson to participants. The participants were randomly divided into two groups (control and experimental). Based on their assignment, they were either presented information in static paragraphs of text centered on the eLearning screen (control), or presented with randomly placed chunks of text on each new eLearning lesson screen. Although the data analysis showed no significant difference in assessment scores for either group, there are suggestions for continued research on this topic. The importance and value of this educational medium requires the use of the best and most impactful methodologies to maximize attention, clarity, learning, and recall. More research needs to be conducted with varying forms of chunking with different populations using eye-tracking to ensure optimized educational endeavors.

[pt] O Processamento de Linguagem natural é uma área de pesquisa que explora como computadores podem entender e manipular textos em linguagem natural. Dentre as tarefas mais conhecidas em PLN está a de rotular sequências de texto. O problema de segmentação de texto em sintagmas é um dos problemas que pode ser abordado como rotulagem de sequências. Para isto, classificamos quais palavras pertencem a um sintagma, onde cada sintagma representa um grupo disjunto de palavras sintaticamente correlacionadas. Este tipo de segmentação possui importantes aplicações em tarefas mais complexas de processamento de linguagem natural, como análise de dependências, tradução automática, anotação de papéis semânticos, identificação de orações e outras. O objetivo deste trabalho é apresentar uma arquitetura de rede neural profunda para o problema de segmentação textual em sintagmas para a língua portuguesa. O corpus usado nos experimentos é o Bosque, do projeto Floresta Sintá(c)tica. Baseado em trabalhos recentes na área, nossa abordagem supera o estado-da-arte para o português ao alcançar um F(beta)=1 de 90,51, que corresponde a um aumento de 2,56 em comparação com o trabalho anterior. Além disso, como forma de comprovar a qualidade do segmentador, usamos os rótulos obtidos pelo nosso sistema como um dos atributos de entrada para a tarefa de análise de dependências. Esses atributos melhoraram a acurácia do analisador em 0,87. / [en] Natural Language Processing is a research field that explores how computers can understand and manipulate natural language texts. Sequence tagging is amongst the most well-known tasks in NLP. Text Chunking is one of the problems that can be approached as a sequence tagging problem. Thus, we classify which words belong to a chunk, where each chunk represents a disjoint group of syntactically correlated words. This type of chunking has important applications in more complex tasks of natural language processing, such as dependency parsing, machine translation, semantic role labeling, clause identification and much more. The goal of this work is to present a deep neural network archtecture for the Portuguese text chunking problem. The corpus used in the experiments is the Bosque, from the Floresta Sintá(c)tica project. Based on recent work in the field, our approach surpass the state-of-the-art for Portuguese by achieving a F(beta)=1 of 90.51, which corresponds to an increase of 2.56 in comparison with the previous work. In addition, in order to attest the chunker effectiveness we use the tags obtained by our system as feature for the depedency parsing task. These features improved the accuracy of the parser by 0.87.