College students are increasingly engaged in watching online videos and using social media. Therefore, researchers should attempt to better understand how these variables as well as other social and cognitive variables, affect the communication competence of students. The researcher administered a questionnaire to 392 college students from a private medium-sized West Coast university using various modified scales examining student-to-student confirmation, social support through social media, online video viewing, self-esteem, personal reports of public speaking anxiety, and communication competence. Student-to-student confirmation and online video viewing were measured using new sets of scales developed by the researcher. The data from the surveys was analyzed to determine which independent variables have the greatest influence on communication competence. Student-to-student confirmation, social support through social media, online video viewing, and self-esteem all have a significant positive relationship with communication competence. The correlation analysis also found a significant negative correlation between personal reports of public speaking anxiety and communication competence. A regression analysis showed that the combination of social support through social media, online video viewing, self-esteem, and personal reports of public speaking anxiety demonstrated predictive power regarding communication competence. These results suggest that there are important relationships between both social and cognitive variables and communication competence.

No description available.

The amount of video content online will nearly triple in quantity by 2021 compared to 2016. The implementation of sophisticated filters is of paramount importance to manage this information flow. The research question of this thesis asks to what extent it is possible to generate personal recommendations, based on the data that news videos implies. The objective is to evaluate how different recommender systems compare to complete random, each other and how they are received by users in a test environment. This study was performed during the spring of 2018, and explore four different algorithms. These recommender systems include a content-based, a collaborative-filter, a hybrid model and a popularity model as a baseline. The dataset originates from a news media startup called Newstag, who provide video news on a global scale. The data is sparse and includes implicit feedback only. Three offline experiments and a user test were performed. The metric that guided the algorithms offline performance was their recall at 5 and 10, due to the fact that the top list of recommended items are of most interest. A comparison was done on different amounts of meta-data included during training. Another test explored respective algorithms performance as the density of the data increased. In the user test, a mean opinion score was calculated based on the quality of recommendations that each of the algorithms generated for the test subjects. The user test also included randomly sampled news videos to compare with as a baseline. The results indicate that for this specific setting and data set, the content-based recommender system performed best in both the recall at five and ten, as well as in the user test. All of the algorithms outperformed the random baseline. / Mängden video som finns tillgänglig på internet förväntas att tredubblas år 2021 jämfört med 2016. Detta innebär ett behov av sofistikerade filter för att kunna hantera detta informationsflöde. Detta examensarbete ämnar att svara på till vilken grad det går att generera personliga rekommendationer baserat på det data som nyhetsvideo innebär. Syftet är att utvärdera och jämföra olika rekommendationssystem och hur de står sig i ett användartest. Studien utfördes under våren 2018 och utvärderar fyra olika algoritmer. Dessa olika rekommendationssystem innefattar tekniker som content-based, collaborative-filter, hybrid och en popularitetsmodell används som basvärde. Det dataset som används är glest och har endast implicita attribut. Tre experiment utförs samt ett användartest. Mätpunkten för algoritmernas prestanda utgjordes av recall at 5 och recall at 10, dvs. att man mäter hur väl algoritmerna lyckas generera värdefulla rekommendationer i en topp-fem respektive topp-10-lista av videoklipp. Detta då det är av intresse att ha de mest relevanta videorna högst upp i sin lista av resultat. En jämförelse gjordes mellan olika mängd metadata som inkluderades vid träning. Ett annat test gick ut på att utforska hur algoritmerna presterar då datasetet blir mindre glest. I användartestet användes en utvärderingsmetod kallad mean-opinion-score och denna räknades ut per algoritm genom att testanvändare gav betyg på respektive rekommendation, baserat på hur intressant videon var för dem. Användartestet inkluderade även slumpmässigt generade videos för att kunna jämföras i form av basvärde. Resultaten indikerar, för detta dataset, att algoritmen content-based presterar bäst både med hänsyn till recall at 5 & 10 samt den totala poängen i användartestet. Alla algoritmer presterade bättre än slumpen.

Several studies have been conducted on the importance of physical activity, especially at a younger age. Participating in physical activity at a younger age often facilitates an active lifestyle in adulthood. Various ways of promoting physical activity have been researched. However, the use of interactive documentaries (i-docs) to promote physical activity has not been investigated. This paper addresses the filming, editing and evaluation of three documentaries relating to the concept of Street Football, a concept created by Svenska Fotbollförbundet (SvFF). The three documentaries contain similar content but are different in level of interactivity. The documentaries were tested in a user study with two test groups of children and young adults between the ages of 11 and 20. The findings show that all the levels of interactivity in the documentaries tested in this project can successfully promote physical activity. The ability to interact and choose what content to see next was especially appreciated by the participants. Further research is required to draw more significant conclusions. / Flera studier har utförts på vikten av fysisk aktivitet, speciellt i en yngre ålder. Deltagande i fysisk aktivitet i en yngre ålder innebär en större sannolikhet för en aktiv livsstil i vuxen ålder. Olika sätt att uppmuntra fysisk aktivitet har undersökts i andra rapporter. Men användningen av interaktiva dokumentärer (i-docs) för att uppmuntra fysisk aktivitet har inte undersökts. Den här rapporten adresserar filmandet, redigerandet och utvärderandet av tre dokumentärer relaterat till konceptet Street Fotboll, ett koncept framtaget av Svenska Fotbollförbundet (SvFF). De tre dokumentärerna innehåller liknande material men skiljer sig i nivån av interaktivitet. Dokumentärerna testades i en användarstudie med två testgrupper som bestod av barn och unga vuxna mellan åldern 11 och 20 år gammal. Upptäckterna visade att alla nivåer av interaktivitet i dokumentärerna som testades i det här projektet kunde framgångsrikt uppmuntra fysisk aktivitet. Förmågan att kunna interagera och välja vad man ville se var speciellt uppskattat av deltagarna. Ytterligare forskning krävs för att dra mer signifikanta slutsatser.

La diferencia entre la televisión de paga y los OTT de contenidos se han ido reduciendo; sin embargo, las diferencias en relación al enfoque regulatorio aplicable se mantienen e incluso se han incrementaron. Así, surge la discusión sobre si estos servicios deben estar sujetos a la misma regulación (simetría regulatoria) o si, por el contrario, corresponde que mantengan una regulación diferenciada (asimetría regulatoria). El objetivo de la investigación es determinar si el servicio de televisión de paga y los OTT de contenidos forman parte de un mismo mercado, lo cual implica determinar su relación de competencia a nivel económico (determinar si son sustitutos). Esto se realizó mediante un modelo Probit, el cual permite identificar el efecto de tener un OTT de contenido sobre la probabilidad de contratar la televisión de paga. Al 2022, la televisión de paga viene reduciendo sus conexiones en servicio en Perú lo cual genera dos efectos sobre este mercado, los efectos de orden primario involucran i) una mayor concentración del mercado de televisión de paga, en algunas regiones, ii) se pone en riesgo el cierre de la brecha digital debido a la salida del mercado de empresas de televisión de paga que brindan internet fijo y iii) el incremento de la informalidad del mercado de televisión de paga mientras que los efectos de orden secundario implican el debilitamiento del Osiptel y del Pronatel debido a un menor aporte por regulación. La investigación estimó la relación entre la televisión de paga y los OTT de contenidos, como biblioteca de películas y/o series, pudiéndose apreciar que la probabilidad de contratar el servicio de televisión de paga se incrementa en 6,4% si el hogar tiene contratado un servicio OTT de contenido. En tal sentido, la “asimetría regulatoria” observada entre ambos servicios no afectaría la dinámica de competencia, por el momento, considerando que estos servicios no forman parte de un mismo mercado (no son bienes sustitutos). La investigación no presenta un análisis empírico sobre la relación entre la televisión de paga y los OTT de contenidos, como proveedor de contenidos en vivo y en directo, debido a la falta de información, aunque presenta un análisis cualitativo, el cual muestra indicios de que exista una relación de sustitución entre estos servicios, siendo necesario futuras investigaciones entre estos dos servicios. / The difference between pay television and OTT content have been narrowing; however, the differences in relation to the applicable regulatory approach remain and have even increased. Thus, the discussion arises about whether these services should be subject to the same regulation (regulatory symmetry) or if, on the contrary, it is appropriate that they maintain a differentiated regulation (regulatory asymmetry). The objective of the investigation is to determine if the pay television service and the OTT content are part of the same market, which implies determining their competitive relationship at an economic level (determining if they are substitutes). This was done using a Probit model, which allows identifying the effect of having a content OTT on the probability of contracting pay television. By 2022, pay television has been reducing its connections in service in Peru, which generates two effects on this market, the effects of a primary order involve i) a greater concentration of the pay television market, in some regions, ii) it becomes The closing of the digital divide is at risk due to the exit from the market of pay television companies that provide fixed internet and iii) the increase in the informality of the pay television market while secondary effects imply the weakening of Osiptel and Pronatel due to a lower contribution due to regulation. The research estimated the relationship between pay television and OTT content, such as a library of films and/or series, being able to appreciate that the probability of contracting the pay television service increases by 6.4% if the household has contracted an OTT content service. In this sense, the "regulatory asymmetry" observed between both services would not affect the dynamics of competition, for the moment, considering that these services are not part of the same market (they are not substitute goods). The research does not present an empirical analysis on the relationship between pay television and OTT content, as a provider of live and direct content, due to the lack of information, although it presents a qualitative analysis, which shows indications that there is a substitution relationship between these services, requiring future research between these two services.

A deficit currently exists in teacher preparation programs for teaching students with Autism Spectrum Disorder (ASD) using evidence-based practices, specifically in the area of Applied Behavior Analysis. As communication deficits are a primary characteristic associated with the diagnosis of ASD teaching communication skills is a common educational goal for students with ASD, and this skill needs to be addressed in teacher education. Mand training is one evidence-based practice applied to teach independent requesting skills to students with ASD. In this study, four teachers serving students with characteristics of ASD in the classroom participated in a multiple-probe across-participants study evaluating the extent to which online training videos prepare teachers in mand training procedures for students with ASD. Concurrent with teacher participants, changes in rates of student mands were evaluated for four students receiving instruction in a multiple probe design. Data analysis consisted of visual analysis of graphically displayed results and calculation of effect sizes. Results indicate that the Online Training Videos (OTV) were not effective as a standalone intervention for preparation of teachers to teach mands, however three student participants showed improved student outcomes.

[pt] Com uma perspectiva estética, filosófica e antropológica, a tese ora apresentada propõe atenção especial à produção audiovisual realizada em instituições formais de ensino, procurando perceber uma imagem dialética da educação a partir de uma seleção de 251 vídeos produzidos em escolas do estado do Rio de Janeiro e exibidos no segmento infanto-juvenil do Festival do Rio, a Mostra Geração, entre os anos de 2000 e 2014. Entendendo a noção benjaminiana da dialética das imagens como dinâmica que conflagra uma experiência, a constituição de um olhar crítico que observa a imagem e que também é afetado por ela, a autora acredita que as imagens nos levam a experimentar um tipo diferenciado de conhecimento. Deste ponto de vista, portanto, surge a relevância de tentar, dentro do contexto acadêmico, uma aproximação sensível (estética) com a escola, para enriquecer as reflexões, a ciência e a imaginação sobre as instituições sociais de formação. Dividida em três fases, a pesquisa teve como principal objetivo testar uma combinação de estratégias da arte, da filosofia e das ciências sociais na produção de conhecimento para o campo da educação. Após uma categorização básica do material empírico com o auxílio de um software de análise quantitativa (ATLAS.ti), a pesquisadora segue um percurso analítico aberto por sua apropriação pessoal dos escritos de Georges Didi-Huberman. A segunda fase, que se desenvolveu concomitantemente às demais, é uma experiência estética/criativa de edição de partes dos audiovisuais escolares, utilizando a montagem como método investigativo e reflexivo para compor um instrumento de consulta que atendesse adois objetivos adicionais: a) compreender e estranhar a imagem de escola partilhada nos vídeos; b) conceber imagens-pensamento com e a partir dos planos que a escola põe em circulação. Em um terceiro momento, a autora produz um discurso como método de desconstrução de um ponto de vista e de (re)constituição de um outro olhar-saber, uma tese, sobre a aparição da escola no material construído. O Atlas do Audiovisual Escolar, Territórios da Mostra Geração: Rio de Janeiro, 2000 a 2014, composto por um filme e um ensaio, reflete paixões e inquietações; emoções e sonhos despertados no encontro com a amostra do imaginário audiovisual escolar abordado. Violência, drogas, corrupção, poluição, mídias e inovação são os problemas que ocupam a imagem de escola formada. Conclui-se que encarar o audiovisual escolar com responsabilidade é, obrigatoriamente, interrogar imagens que nos olham e imaginar, com as escolas, novas imagens por vir. / [en] With an aesthetic, philosophical and anthropological perspective, her doctoral research proposes special attention to the moving imagens of the video production carried out in educational institutions, analyzing a selection of 251 videos produced by schools in the state of Rio de Janeiro and exhibited in the Generation Programme - the Rio Int l Film Festival s Segment for Children and Young People, between the years 2000 and 2014. The study starts from the Benjaminian concept of dialectic images understooded as a dynamic that constitutes a critical gaze that look to an image while the observer is also looked by the image (Benjamin). Believing that images can lead us to experience a peculiar kind of knowledge, emerges the relevance of attempting, within the academic context, a sensitive (aesthetic) approach with the school, to contribute to reflections, to the science and the imagination on the educational institutions. Divided into three phases, the main objective of the research is to test a combination of strategies of art, philosophy and social sciences in the construction of knowledge for the educational research field. The first approach of the data collected was a categorization of the empirical material with the aid of quantitative analysis software (ATLAS.ti). Then the researcher follows an analytical path opened by her personal appropriation of the writings of Georges Didi-Huberman. The second phase, which develops concomitantly to the others, is an aesthetic / creative experience of editing parts of the school-produced videos, using montage and remix as investigative and reflexive methods to compose a query instrument that serves two additional objectives: a) to compreheend and to strange the image of school that is shared by the videos of this study; b) to conceive Thought-Images (Denkbilder) with and from the frames that the schools put into circulation. In a third moment, the author discuss what was visualized, by producing a discourse as a method of deconstructing a point of view and (re) constitution of another gazeknowledge, a thesis about the school s appearance in the builded material. The Atlas of the School-produced Video: territories of Generation Programme/Rio de Janeiro Int l Film Festival (2000-2014), composed of a film and an essay, reflects passions and anxieties; emotions and dreams awakened in the encounter with the sample of the studied video-imaginary. Violence; Drugs; Corruption; Pollution; Media and Innovation are issues that occupy the school image formed in the remix produced by the research. It is concluded that video remix is a methodology that supports creation of knowledge in both Media Education and Educational fields, since it is a way to manipulate and, hence, is a form of material engagement with the data. Finally, the researcher states that to see school-produced videos, with responsibility in fact, necessarily is to ask questions to the images that affect us and to imagine with the schools new images to come.

Syftet med uppsatsen är att undersöka hur skogen kan användas för att gestalta en bild avidentiteten. För att göra det analyseras musikvideor av Fever Ray och Iamamiwhoami.Gestaltningen består av två noveller. I dessa undersöks relationen till skogen på ettmer individuellt plan. Metoden är mer sökande än teoretiserande. Uppsatsen analyserarmusikvideor ur ett genusperspektiv, som bilder och metaforer. Teoretiker som används ärbland annat Jean Baudrillard, Timothy Morton och Judith Butler. Från deras teorier undersöksvilka värderingar och antaganden som kan sägas ligga till grund för representationen avidentiteten och skogen. Resultatet är bilden av en värld som är splittrad och ovillig att kommatill konkreta ståndpunkter. Å ena sidan visar musikvideorna individer som arbetar medkulturellt betingade bilder av genus och skogen och försöker göra dem till sina egna. Å andrasidan syns en vilja att röra sig bort från dem. Likaså porträtterar gestaltningen personer somkämpar för att komma tillrätta med sin plats i den givna ordningen men aldrig riktigt lyckas. / The purpose of this essay is to investigate how the forest can be used to mirror the identity. Indoing so, music videos by Fever Ray and Iamamiwhoami are being analyzed. The portrayalconsists of two short stories. In these, the relation to the forest is investigated on a moreindividual level. The method is more searching than theorizing. The essay analyzes musicvideos, with a gender perspective, as images and metaphors. Theorists used are amongothers Jean Baudrillard, Timothy Morton and Judith Butler. From their theories the thesisinvestigates the values and assumptions that can be said to constitute the representation of theforest and the identity in the videos. The result is the image of a fragmented world, unwillingto come to concrete conclusions. On one hand the music videos show individuals who workswith the overtaking of given cultural images of the forest and gender. On the other hand a willto reject the same images can be seen. Likewise, the portrayal shows individuals who struggleto come to terms with their place in the given order but never really succeeds.

This study investigated the ways in which college algebra students watch mathematics instructional videos with the goal of answering the following research questions: (1) How are student responsibility frames similar or different among students within a particular video? (2) How do the various video design principles support or constrain the uses of particular responsibility frames? This research was guided by the cognitive theory of multimedia design and the theory of didactic situations. The cognitive theory of multimedia design outlined principles for video creation and design that could impact student learning from video watching. The theory of didactic situations defined implicit teacher and student responsibilities within the context of the face-to-face mathematics classroom and was applied in this study to students watching mathematics instructional videos. Participants were recruited from five college algebra classes at a university in the northeastern United States and were asked to attend two semi-structured interviews. During the first interview, participants were pretested to determine their prior knowledge about how to solve quadratic functions and to measure their mathematics-related beliefs. During the second interview, participants watched three different videos about solving quadratic equations by completing the square and were asked questions about the mathematics that they viewed in the videos. Transcriptions of audio recordings of these interviews were coded thematically using categories previously identified by the didactic contract for the face-to-face classroom and expanding the types of student responsibilities identified specifically for video watching as needed. This study found that participants, regardless of overall prior knowledge or mathematics-related beliefs, but who had prior knowledge of completing the square, held a responsibility to use the specific set of steps they were taught by their teacher to solve problems. Additionally, participants at some levels of prior knowledge expressed a responsibility to not watch videos that showed a visual representation of the mathematics. Findings also suggested a student heuristic for the selection and use of mathematics instructional videos that may be useful to both mathematics teachers and video creators. / Math & Science Education

<p dir="ltr">Extensive data availability has catalyzed the expansion of deep learning. Such advancements include image classification, speech, and natural language processing. However, this data-driven progress is often hindered by privacy restrictions preventing the public release of specific datasets. For example, some vision datasets cannot be shared due to privacy regulations, particularly those containing images depicting visually sensitive or disturbing content. At the same time, it is imperative to deploy deep learning efficiently, specifically Deep Neural Networks (DNNs), which are the core of deep learning. In this dissertation, we focus on achieving efficiency by reducing the computational cost of DNNs in multiple ways.</p><p dir="ltr">This thesis first tackles the privacy concerns arising from deep learning. It introduces a novel methodology that synthesizes and releases synthetic data, instead of private data. Specifically, we propose Differentially Private Image Synthesis (DP-ImgSyn) for generating and releasing synthetic images used for image classification tasks. These synthetic images satisfy the following three properties: (1) they have DP guarantees, (2) they preserve the utility of private images, ensuring that models trained using synthetic images result in comparable accuracy to those trained on private data, and (3) they are visually dissimilar from private images. The DP-ImgSyn framework consists of the following steps: firstly, a teacher model is trained on private images using a DP training algorithm. Subsequently, public images are used for initializing synthetic images, which are optimized in order to be aligned with the private dataset. This optimization leverages the teacher network's batch normalization layer statistics (mean, standard deviation) to inject information from the private dataset into the synthetic images. Third, the synthetic images and their soft labels obtained from the teacher model are released and can be employed for neural network training in image classification tasks.</p><p dir="ltr">As a second direction, this thesis delves into achieving efficiency in deep learning. With neural networks widely deployed for tackling diverse and complex problems, the resulting models often become parameter-heavy, demanding substantial computational resources for deployment. To address this challenge, we focus on quantizing the weights and the activations of DNNs. In more detail, we propose a method for compressing neural networks through layer-wise mixed-precision quantization. Determining the optimal bit widths for each layer is a non-trivial task, given the fact that the search space is exponential. Thus, we employ a Multi-Layer Perceptron (MLP) trained to determine the suitable bit-width for each layer. The Kullback-Leibler (KL) divergence of softmax outputs between the quantized and full precision networks is the metric used to gauge quantization quality. We experimentally investigate the relationship between KL divergence and network size, noting that more aggressive quantization correlates with higher divergence and vice versa. The MLP is trained using the layer-wise bit widths as labels and their corresponding KL divergence as inputs. To generate the training set, pairs of layer-wise bit widths and their respective KL divergence values are obtained through Monte Carlo sampling of the search space. This approach aims to reduce the computational cost of DNN deployment, while maintaining high classification accuracy.</p><p dir="ltr">Additionally, we aim to enhance efficiency in machine learning by introducing a computationally efficient method for action recognition on compressed videos. Rather than decompressing videos for action recognition tasks, our approach performs action recognition directly on the compressed videos. This is achieved by leveraging the modalities within the compressed video format, specifically motion vectors, residuals, and intra-frames. To process each modality, we deploy three neural networks. Our observations indicate a hierarchy in convergence behavior: the network processing intra-frames tend to converge to a flatter minimum than the network processing residuals, which, in turn, converge to a flatter minimum than the motion vector network. This hierarchy motivates our strategy for knowledge transfer among modalities to achieve flatter minima, generally associated with better generalization. Based on this insight, we propose Progressive Knowledge Distillation (PKD), a technique that incrementally transfers knowledge across modalities. This method involves attaching early exits, known as Internal Classifiers (ICs), to the three networks. PKD begins by distilling knowledge from the motion vector network, then the residual network, and finally the intra-frame network, sequentially improving the accuracy of the ICs. Moreover, we introduce Weighted Inference with Scaled Ensemble (WISE), which combines outputs from the ICs using learned weights, thereby boosting accuracy during inference. The combination of PKD and WISE demonstrates significant improvements in efficiency and accuracy for action recognition on compressed videos.</p><p dir="ltr">In summary, this dissertation contributes to advancing privacy preserving and efficient machine learning algorithms. The proposed methodologies offer practical solutions for deploying machine learning systems in real-world scenarios by addressing data privacy and computational efficiency. Through innovative approaches to image synthesis, neural network compression, and action recognition, this work aims to foster the development of robust and scalable machine learning frameworks for diverse computer vision applications.</p>