Dissertations / Theses on the topic 'Processamento Genere'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 15 dissertations / theses for your research on the topic 'Processamento Genere.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
MORNATI, GIULIA. "THE SENSITIVITY AND PRODUCTION OF ARTICLES AND 3rd PERSON DIRECT OBJECT CLITICS: EVIDENCE FROM EYE-MOVEMENTS, ERP RECORDINGS AND ONLINE TESTS IN ITALIAN CHILDREN AND TODDLERS." Doctoral thesis, Università degli Studi di Milano-Bicocca, 2022. http://hdl.handle.net/10281/379211.
Full textFunction words and inflectional morphemes do not have an independent meaning but are highly frequent and usually monosyllabic. Among these, we could find articles, clitic pronouns, nominal and verbal suffixes. They convey structural information and are a critical aid in language acquisition and processing. Although previous research has demonstrated that infants are sensitive to function words from their first months of life, studies on their sensitivity and production in Italian are scarce. This thesis aims to study the sensitivity and production of Italian infants and toddlers to articles and 3rd direct object (3DO) clitics. This choice is motivated by the fact that these two morphemes are critical in the detection of language problems, especially 3DO, in monolingual and bilingual children. Chapter 3 investigated the early sensitivity to articles and 3DO clitics. By recording eye movements, we analysed whether 12 months old infants distinguished between real and pseudo-articles and whether 20 months old toddlers distinguished between real and pseudo-clitics. Results showed that, concerning the articles, infants distinguished between real and pseudo articles, preferring to look at the screen when they heard a real article. Moreover, at already 12 months of age, Italian infants had acquired both definite and indefinite articles. No significant results emerged with 3DO clitics due, probably, to the methodology used. Chapters 4 and 5 deeply investigated the use of an inflection morpheme, gender, on articles in language processing. Through a Looking While Listening task, we showed that infants as young as 12 months could extract the gender information carried by articles to anticipate the upcoming noun (Chapter 4). Finally, ERP recordings showed that children as young as 24 months old could detect a gender violation between a picture and the associated sentence already when they heard that the article did not match the picture presented: they showed a very early posterior positivity (Chapter 5). Finally, we studied the production of 3DO clitics by presenting a helpful computerised test for screening the production of 3DO clitics both in the typical developing (TD) children and in children with Developmental Language Disorder (DLD). The test had a good sensitivity in identifying children with DLD. Moreover, this test showed interesting results in the bilingual population: although children who acquire Italian as a second language struggle with the acquisition of 3DO clitics, TD bilingual children performed as well as monolingual peers at the clitic test (Chapter 6). Finally, we have deepened the study of 3DO clitics’ production by analysing the differences in the production of proclitics and enclitics in children aged 4 to 7 years with and without familiarity with language and learning impairments (LLI). Results showed that enclitics were easier to produce at all ages. Moreover, children with a positive family history for LLI produced fewer clitics than children without familiarity, especially at 4 and 5 years.
MIOLLA, ALESSIO. "PROCESSAMENTO E RILEVAMENTO DI ESPRESSIONI FACCIALI EMOTIVE SPONTANEE E SIMULATE." Doctoral thesis, Università degli studi di Padova, 2022. http://hdl.handle.net/11577/3459406.
Full textFacial expressions are the most effective and reliable indicator of emotional states. However, people are adept at modulating and falsifying their emotional expressions according to their needs, completely changing the observer’s perception and reaction. A new validated dataset displaying more than 1450 clips of both spontaneous and posed emotional facial expressions was created for a dual purpose: discriminate spontaneous and posed emotions according to the facial movements (detection section) and investigate how the brain extracts the genuineness of emotional expressions (perception section). In the detection part, Machine Learning models were applied to the clips in order to discriminate spontaneous from posed emotion automatically. Results yielded high accuracies in genuineness discrimination (up to 84.4% accuracy). Moreover, for the first time, the 3-D motion analysis was applied to the study of spontaneous and posed dynamic facial expressions of happiness to detect subtle movements in terms of space, time, and speed. Results revealed that the mouth widening and the speed of smiles are greater in posed than spontaneous happiness. In the perception section, time-frequency EEG analysis was used to compare the perception of three spontaneous and posed emotional facial expressions: happiness, disgust, and fear. Overall, strong differences in both the timing and the topography of the canonical EEG bands were observed, revealing how spontaneous and posed emotions are processed differently in our brains. In particular, compared to genuine happiness, posed happiness revealed increased delta and theta power at the onset and offset of the facial expressions over frontal sites. Compared to posed fear, genuine fear elicits an increase in alpha and beta bands followed by an increase in theta activity. Finally, for facial expressions of disgust, we found an early increased theta, alpha, and beta activity for the posed expressions, followed by increased activity in alpha and beta bands during the perception of genuine disgust. The implications and applications of these studies are discussed in light of the state of art of lie detection, psychology of emotions, and the AI field.
Martins, Guilherme Brandão [UNESP]. "Processamento e análise de vídeos utilizando Floresta de Caminhos Ótimos." Universidade Estadual Paulista (UNESP), 2016. http://hdl.handle.net/11449/139458.
Full textApproved for entry into archive by Ana Paula Grisoto (grisotoana@reitoria.unesp.br) on 2016-06-13T17:06:19Z (GMT) No. of bitstreams: 1 martins_gb_me_sjrp.pdf: 11362535 bytes, checksum: c1da2ab3e80ead0846eae49d9a1bc40e (MD5)
Made available in DSpace on 2016-06-13T17:06:19Z (GMT). No. of bitstreams: 1 martins_gb_me_sjrp.pdf: 11362535 bytes, checksum: c1da2ab3e80ead0846eae49d9a1bc40e (MD5) Previous issue date: 2016-05-20
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
Com os avanços relacionados às tecnologias de redes computacionais e armazenamento de dados observa-se que, atualmente, uma grande quantidade de conteúdo digital está sendo disponibilizada via internet, em especial por meio de redes sociais. A fim de explorar esse contexto, abordagens relacionadas ao processamento e apredizado de padrões em vídeos têm recebido crescente atenção nos últimos anos. Sistemas de recomendação de filmes, amplamente empregados em lojas virtuais, são umas das principais aplicações no que se refere aos avanços de pesquisa na área de processamento de vídeos. Com o objetivo de acelerar o processo de recomendação e redução de armazenamento, técnicas para classificação e sumarização de vídeos por meio de aprendizado de máquina têm sido utilizadas com o intuito de explorar conteúdo informativo e também redundante. Por meio de técnicas de agrupamento e descrição de dados, é possível identificar quadros-chave de um conjunto de amostras a fim de que, posteriormente, estes sejam usados para sumarização do vídeo. Além disso, por meio de bases de vídeos rotuladas, podemos classificar amostras de modo a organizá-las por gêneros de vídeo. O presente trabalho objetiva utilizar o classificador Floresta de Caminhos Ótimos para sumarização automática e classificação de vídeos por gênero, bem como o estudo de sua viabilidade nestes contextos. Os resultados obtidos mostram que o referido classificador obteve desempenhos bastante promissores e próximos à algumas das técnicas de sumarização automática e classificação de vídeos que, atualmente, representam o estado-da-arte no atual contexto.
Currently, a number of improvements related to computational networks and data storage technologies have allowed a considerable amount of digital content to be provided on the internet, mainly through social networks. In order to exploit this context, video processing and pattern recognition approaches have received a considerable attention in the last years. Movie recommendation systems are widely employed in virtual stores, thus being one of the main applications regarding to research advances in the video processing field. Aiming to boost the content recommendation and storage cutback, different video categorization and video summarization techniques have been applied to handle with more informative and redundant content. By availing clustering and data description techniques, it is possible to identify keyframes from a given sample collection in order to consider them as part of the video summarization process. Furthermore, through labeled video data collections it is possible to classify samples in order to arrange them by video genres. The main goal of this work is to employ the Optimum-Path Forest classifier in both video summarization and video genre classification processes as well as to conduct a viability study of such classifier in the aforementioned contexts. The results have shown this classifier can achieve promising performances, being very close in terms of summary quality and consistent recognition rates to some state-of-the-art video summarization and classification approaches.
Azevedo, Bernardo Lopes de Sá. "Reconstrução/processamento de imagem médica com GPU em tomossíntese." Master's thesis, Faculdade de Ciências e Tecnologia, 2011. http://hdl.handle.net/10362/7503.
Full textA Tomossíntese Digital Mamária (DBT) é uma recente técnica de imagem médica tridimensional baseada na mamografia digital que permite uma melhor observação dos tecidos sobrepostos, principalmente em mamas densas. Esta técnica consiste na obtenção de múltiplas imagens (cortes) do volume a reconstruir, permitindo dessa forma um diagnóstico mais eficaz, uma vez que os vários tecidos não se encontram sobrepostos numa imagem 2D. Os algoritmos de reconstrução de imagem usados em DBT são bastante similares aos usados em Tomografia Computorizada (TC). Existem duas classes de algoritmos de reconstrução de imagem: analíticos e iterativos. No âmbito deste trabalho foram implementados dois algoritmos iterativos de reconstrução: Maximum Likelihood – Expectation Maximization (ML-EM) e Ordered Subsets – Expectation Maximization (OS-EM). Os algoritmos iterativos permitem melhores resultados, no entanto são computacionalmente muito pesados, pelo que, os algoritmos analíticos têm sido preferencialmente usados em prática clínica. Com os avanços tecnológicos na área dos computadores, já é possível diminuir consideravelmente o tempo que leva para reconstruir uma imagem com um algoritmo iterativo. Os algoritmos foram implementados com recurso à programação em placas gráficas − General-Purpose computing on Graphics Processing Units (GPGPU). A utilização desta técnica permite usar uma placa gráfica (GPU – Graphics Processing Unit) para processar tarefas habitualmente designadas para o processador de um computador (CPU – Central Processing Unit) ao invés da habitual tarefa do processamento gráfico a que são associadas as GPUs. Para este projecto foi usado uma GPU NVIDIA®, recorrendo-se à arquitectura Compute Unified Device Architecture (CUDA™) para codificar os algoritmos de reconstrução. Os resultados mostraram que a implementação dos algoritmos em GPU permitiu uma diminuição do tempo de reconstrução em, aproximadamente, 6,2 vezes relativamente ao tempo obtido em CPU. No respeitante à qualidade de imagem, a GPU conseguiu atingir um nível de detalhe similar às imagens da CPU, apesar de diferenças pouco significativas.
Perri, Juliana Miguel. "Bacterias do genero Enterococcus em queijo de coalho = influencia do processamento na seleção microbiana, perfil tecnologico e implicações em segurança de alimentos." [s.n.], 2010. http://repositorio.unicamp.br/jspui/handle/REPOSIP/255488.
Full textDissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia de Alimentos
Made available in DSpace on 2018-08-15T04:42:43Z (GMT). No. of bitstreams: 1 Perri_JulianaMiguel_M.pdf: 449133 bytes, checksum: cbe24cd7c7343f2e8c5bead5fae83aa2 (MD5) Previous issue date: 2010
Resumo: Bactérias do gênero Enterococcus ocorrem naturalmente como microbiota lática numa variedade de queijos. Esta presença pode ser relacionada à contaminação, porém pode ter caráter positivo devido à propriedades tecnológicas descritas para o gênero. Este trabalho teve como objetivo avaliar a influência do processamento de queijo de coalho na seleção de bactérias do gênero Enterococcus, como também determinar o perfil tecnológico e as possíveis implicações em segurança de alimentos da presença de bactérias desse gênero no produto queijo de coalho. Para elaboração dos queijos, os parâmetros de tratamento térmico foram: pasteurização rápida e cozimento da massa em três diferentes temperaturas (40, 45 e 50°C). Os resultados evidenciaram a ação selet iva do processamento para bactérias do gênero Enterococcus, sendo a etapa de pasteurização incapaz da eliminação. Tanto a etapa de cozimento da massa, quanto de salga favoreceram o crescimento deste gênero de modo seletivo dentre o total de bactérias láticas presentes, sem discriminação entre as diferentes temperaturas de cozimento utilizadas. Dentre as espécies identificadas, E. faecium foi a de maior incidência (83,9%), seguida pela espécie E. faecalis (0,6%). O perfil tecnológico das bactérias identificadas mostrou que estas podem ser capazes de contribuir ativamente com o desenvolvimento de atributos de sabor e aroma característicos ao produto. As bactérias do gênero Enterococcus isoladas apresentaram baixa atividade acidificante porém há incidência de isolados com características de fermento termofílico que podem participar na obtenção de resultados tecnológicos satisfatórios neste produto queijo de coalho. A presença destas bactérias em alimentos deve ser avaliada principalmente em relação à patogenicidade relacionada à gelatinase, encontrada em apenas uma cepa, e a outras características não detectadas, mas capazes de serem adquiridas
Abstract: Bacteria of the genus Enterococcus occur naturally as a lactic microbiota in a variety of chesses. This presence may be related to contamination however may have positive character due to the technological properties described for the genus. This work aimed to evaluate the influence of coalho cheese processing in the selection of bacteria of the genus Enterococcus as well as determine the technological profile and possible food safety implications by the presence of bacteria in the coalho cheese product. For cheese manufacturing, the parameters of heat treatment were: rapid pasteurization and curd cooking at three different temperatures (40, 45 and 50°C). The results showed the selective action of the processing for bacterias of the genus Enterococcus, having, the step of pasteurization, been incapable of elimination. Both the steps of curd cooking and the curd salting selectively favored the growth of this genus, within the total of the present lactic acid bacteria, without any discrimination between the different temperatures used. Within the bacteria species identified, E. faecium was the most occurrent (83,9%), followed by the specie E. faecalis (0,6%). The technological profile of the identified bacteria showed that they may be able to actively contribute to the development of the characteristic flavour to the product. The isolated bacteria of the genus Enterococcus showed low acidifying activity but there are incidence of isolates with thermofilic starter characteristics that may be able to participate in obtaining satisfactory results in this coalho cheese product. The presence of these bacteria in foods must be evaluated mainly due to the characteristic of pathogenicity-related gelatinase, found in only one isolated strain, and other features not found but that could be acquired
Mestrado
Mestre em Tecnologia de Alimentos
Koetz, Clara Isabel. "A influência do afeto e do gênero do consumidor no processamento das informações de qualidade sinalizadas por meio da propaganda." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2011. http://hdl.handle.net/10183/31967.
Full textThe objective of this thesis is to examine the influence of affect and gender on consumer perceptions of product quality in the signaling quality processes through advertising. In order to do that, two experimental studies were conducted. The first study was implemented in Brazil, in a sample of 144 students, whose objective was to examine the influence of the amount invested in advertising (below the average, on the average and above the average invested by the main competitors), the affect (positive and negative) and the gender (male and female) in the perception of product quality. In the data analysis, two ANOVA models were conducted. In the first one, the dependent variable "Overall Quality” was applied, which refers to a more objective assessment of the product. In the second model, the dependent variable "Perceived Quality" was used, which considers a more subjective evaluation of the product. The second study was developed in France, in a sample of 401 students, whose goal was to examine the influence of affect (positive and negative), the amount invested in advertising (below average, on the average and above the average invested by the main competitors), product information (with and without product information) and the subject's gender (male and female) in the perception of product quality in the signaling quality processes through advertising. The data analysis was developed in four steps. First, an ANOVA was conducted with the independent variables "Affect", "Information" and "Advertising Budget", and the same dependent variables that were used to evaluate the product quality perception in the previous study (in two different models). Next, it was developed another ANOVA model, in which it was defined as independent variables the "Advertising Budget ", "Gender" and "Information", and as dependent variables the same variables of quality perception used before (again in two distinct models). The results showed an interaction between the presence of information and the consumer affective state in the product quality perception. Thus, in situations where the respondents had access to product information, they considered this information in the evaluation of the product. In such cases, the affective state did not influence the consumer perceptions of product quality. However, when information was not provided, the product evalua tion was influenced by the affective state of the respondents. Thus, those who were experiencing positive affective states evaluated better the product than those who were experiencing negative emotional states. Moreover, the results showed an interaction between the gender and the amount invested in advertising in the product quality judgment. In the contexts of low investment in advertising, where the signal quality has not been used, women evaluated the product better than men. However, in the contexts where investments in the market’s average and above the market’s average were applied, in which the quality signal was transmitted, men and women showed similar mean values of quality perception only when the subjective dependent variable of perceived qua lity was used. Thus, it can be considered that the quality signal established a heuristic signal, which was considered more by men than by women in the evaluation of the product quality.
Teigão, Roberta Geneci Neves Weber. "Análise de vídeo por ritmo visual e morfologia em cores / Roberta Geneci Neves Weber Teigão ; orientador, Jacques Facon." reponame:Biblioteca Digital de Teses e Dissertações da PUC_PR, 2007. http://www.biblioteca.pucpr.br/tede/tde_busca/arquivo.php?codArquivo=1409.
Full textBibliografia: f. 88-92
Com o avanço da tecnologia digital e o consequente crescimento na utilização de vídeos digitais, aumenta-se a necessidade de recuperação de informação de interesse nesta mídia que apresenta enormes volumes de dados. Muitas pesquisas sobre indexação e proc
With the advance of digital technology and consequent increase in the use of the digital videos, the necessity for recovering interesting information in this enormous volume of data increases. Many researches on digital video indexing and processing have
Lima, Adriano Alves de. "Classificação automática de gêneros musicais em bases de letras de músicas em diferentes idiomas utilizando métodos de combinação de classificadores." Universidade Tecnológica Federal do Paraná, 2015. http://repositorio.utfpr.edu.br/jspui/handle/1/2553.
Full textThe automatic music genre classification is an important area in music information retrieval. For assist this task many languages resources of natural language processing has been used in song lyrics. This is important considering that the classification audio-based might be comple- mented with lyrics features to improve the results. This work analyzes such natural language processing tools employed to song lyrics in different languages for the music genre classi- fication. Furthermore, we studied other techniques to improve the classification results, for example, features combination by mean of early and late fusion methods. To run the experiments were created three lyrics database in various music genres into the Latin languages (Portuguese and Spanish), Nordic (Danish, Norwegian and Swedish) and En- glish. The results obtained shown that the combination approaches by mean of late fusion with product of the probability and early fusion outperformed the other approaches in music genre classification using song lyrics in different languages.
Piera, Fernández Jaume. "Signal processing of microstructure profiles: integrating turbulent spatial scales in aquatic ecological modeling." Doctoral thesis, Universitat de Girona, 2002. http://hdl.handle.net/10803/7650.
Full textThe answers to these questions were the final specific objectives of this thesis.
Spizzirri, Giancarlo. "Morfometria cerebral e imagens de tensores de difusão da microestrutura de substância branca em homens para mulheres transexuais antes e durante o processo transexualizador." Universidade de São Paulo, 2016. http://www.teses.usp.br/teses/disponiveis/5/5142/tde-18112016-123020/.
Full textINTRODUCTION: Gender Dysphoria (GD) is the incongruence between sex at birth versus its perception and manifestation through individual behavior, and the suffering it is accompanied by. As for transsexuality, it has been suggested that the brain does not follow the same pattern as the rest of the body, since evidence has shown that certain brain regions in transsexual women (TW) - male biological sex - are similar to those in women with no GD. MRI-based studies have investigated brain morphometry through voxel based technique (VBM) and diffusion tensor imaging (DTI) to evaluate the microstructure of white matter (WM). Those studies have verified a number of regions with gray matter (GM) and white matter (WM) volume differences, as well as changes in the microstructure of WM through DTI between genders. OBJECTIVES: to investigate GM and WM volume changes through VBM, as well as the microstructure of WM through DTI in four groups of individuals. METHODS: A sample of 80 individuals was analyzed, as follows: 20 TW with GD not on sexual steroids, and 20 TW with GD on sexual steroids for at least one year (TWh), 20 men and 20 women (control groups). All were submitted to MRI in 1.5T equipment. Two image sequences were obtained from that exam: (1) T1 weighted images for VBM analysis in compliance with DARTEL protocol; and (2) diffusion weighted imaging through TBSS (Tract-Based Spatial Statistics) to obtain fractional anisotropy (FA). In both methods statistical analysis was conducted through Statistical Parametric Mapping (SPM), Version 8. Comparisons between groups were conducted through covariance analysis (ANCOVA), available on SPM. Post-hoc tests were then applied to investigate volume differences. Regions of interest were analyzed after having been a priori defined through morphometric studies and DTI for sexual dimorphism and research work with TW. A priori presupposed, as well as non-predictable, results were reported to resist the statistical threshold of p<=0.05, corrected for multiple comparisons (Family Wise Error/FWE). Another criterion was that regions with at least 30 voxels would be mentioned. RESULTS: VBM showed: (i) GM volume increase in Brodmann 6 area in TWh; (ii) regions with GM volume decrease in the insula and in both brain hemispheres in TW; (iii) genu of corpus callosum volume increase in individuals who share female identity gender. DTI showed: FA lower values in right superior longitudinal fasciculus (SLF) and in the body of corpus callosum of TWh. CONCLUSION: significant regional brain volume differences assessed by VBM were detected in TW groups relative to controls. In the TWh group, brain volume changes suggest a possible influence of cross-sex hormone theraphy on brain neuroplasticity. There are some indication of feminization in the brains of TW individuals, most evidently in the CC and more subtly in the posterior superior frontal GM. The pattern of GM volumes alterations in the insula in the two independent TW groups are novel; such finding may be a marker of transsexuality or be a brain correlate of the suffering associated with this condition. FA value reduction at right SLF has been suggested to be associated to GD
Riccio, Edson Luiz. "Uma contribuição ao estudo da contabilidade como sistema de informação." Universidade de São Paulo, 1992. http://www.teses.usp.br/teses/disponiveis/12/12131/tde-19012009-121736/.
Full textThe continuous release of new concepts in the development and control of an Organization, the advances of Managerial Accounting plus the growing requirements for Financial Accounting present all together a challenge for Accounting and for all individuals somehow involved in its usage, operation, research, development and implementation. This challenge has been stronger due to the increase in the complexity of an organization and the impact caused by the use of Computers, which have introduced the concept of speed in business. Additionally, the influence of the Japanese culture in Business Administration, more specifically, the Kanban, Just in Time, Quality Circles and flexibility, has caused many changes in the organizations behavior, pressuring the Accounting area to act accordingly. Several authors have questioned about these points as for example: Bromwich & Hopwood in Research and Current issues in Management Accounting (1986), saying: Different approaches to organizational structuring are creating different information needs for decision making and control. Greater emphasis is being placed on the need to service both local and central management needs for information, and more, Accounting is also being called upon to play a quite explicit role in the mobilization of organizational change. No longer seen as a mere passive reckoning of the economics of the past, accounting systems are now being implicated in the design of different approaches to the management of organizational visibility and different incentives for organizational action. This aspect had also been identified by Iudícibus in Teoria da Contabilidade (1980) when discussing the Categories of accounting information users and their objectives. Thus, the Accounting and the Accountants are being requested more and more to supply the entire organization with better service and information. The Accountant, as a professional, is asked to be a businessman, changing from his old position of data owner to a new one, as an Information Service Agent, and as a decision making facilitator. Therefore, the Accountant and the Accounting System must be fully integrated with the companys objectives internally and externally. The question is, how the Accounting System can satisfy the information requirements of a Company. The objective of this study is to revise the systems approach for Accounting in a computerized environment considering the new concepts above explained.
Testolin, Alberto. "Modeling cognition with generative neural networks: The case of orthographic processing." Doctoral thesis, Università degli studi di Padova, 2015. http://hdl.handle.net/11577/3424619.
Full textIn questa tesi vengono studiati alcuni processi cognitivi utilizzando recenti modelli di reti neurali generative. A differenza della maggior parte dei modelli connessionisti, l’approccio computazionale adottato in questa tesi enfatizza la natura generativa della cognizione, suggerendo che uno degli obiettivi principali dei sistemi cognitivi sia quello di apprendere un modello interno dell’ambiente circostante, che può essere usato per inferire relazioni causali ed effettuare previsioni riguardo all’informazione sensoriale in arrivo. In particolare, viene considerata una potente classe di reti neurali ricorrenti in grado di apprendere modelli generativi probabilistici dall’esperienza, estraendo informazione statistica di ordine superiore da un insieme di variabili in modo totalmente non supervisionato. Questo tipo di reti può essere formalizzato utilizzando la teoria dei modelli grafici probabilistici, che consente di descrivere con lo stesso linguaggio formale sia modelli di reti neurali che modelli Bayesiani strutturati. Inoltre, architetture di rete di base possono essere estese per creare sistemi più sofisticati, sfruttando molteplici livelli di processamento per apprendere modelli generativi gerarchici o sfruttando connessioni ricorrenti direzionate per processare informazione organizzata in sequenze. Riteniamo che queste architetture avanzate costituiscano un’alternativa promettente alle più tradizionali reti neurali supervisionate di tipo feed-forward, perché riproducono più fedelmente l’organizzazione funzionale e strutturale dei circuiti corticali, consentendo di spiegare come l’evidenza sensoriale possa essere effettivamente combinata con informazione contestuale proveniente da connessioni di feedback (“top-down”). Per giustificare l’utilizzo di questo tipo di modelli, in una serie di simulazioni studiamo nel dettaglio come implementazioni efficienti di reti generative gerarchiche e temporali possano estrarre informazione da grandi basi di dati, contenenti migliaia di esempi di training. In particolare, forniamo evidenza empirica relativa al riconoscimento di caratteri stampati e manoscritti appartenenti a diversi sistemi di scrittura, che possono in seguito essere combinati spazialmente o temporalmente per costruire unità ortografiche più complesse come quelle rappresentate dalle parole inglesi.
Salamon, Justin J. "Melody extraction from polyphonic music signals." Doctoral thesis, Universitat Pompeu Fabra, 2013. http://hdl.handle.net/10803/123777.
Full textLa industria de la música fue una de las primeras en verse completamente reestructurada por los avances de la tecnología digital, y hoy en día tenemos acceso a miles de canciones almacenadas en nuestros dispositivos móviles y a millones más a través de servicios en la nube. Dada esta inmensa cantidad de música al nuestro alcance, necesitamos nuevas maneras de describir, indexar, buscar e interactuar con el contenido musical. Esta tesis se centra en una tecnología que abre las puertas a nuevas aplicaciones en este área: la extracción automática de la melodía a partir de una grabación musical polifónica. Mientras que identificar la melodía de una pieza es algo que los humanos pueden hacer relativamente bien, hacerlo de forma automática presenta mucha complejidad, ya que requiere combinar conocimiento de procesado de señal, acústica, aprendizaje automático y percepción sonora. Esta tarea se conoce en el ámbito de investigación como “extracción de melodía”, y consiste técnicamente en estimar la secuencia de alturas correspondiente a la melodía predominante de una pieza musical a partir del análisis de la señal de audio. Esta tesis presenta un método innovador para la extracción de la melodía basado en el seguimiento y caracterización de contornos tonales. En la tesis, mostramos cómo se pueden explotar las características de contornos en combinación con reglas basadas en la percepción auditiva, para identificar la melodía a partir de todo el contenido tonal de una grabación, tanto de manera heurística como a través de modelos aprendidos automáticamente. A través de una iniciativa internacional de evaluación comparativa de algoritmos, comprobamos además que el método propuesto obtiene resultados punteros. De hecho, logra la precisión más alta de todos los algoritmos que han participado en la iniciativa hasta la fecha. Además, la tesis demuestra la utilidad de nuestro método en diversas aplicaciones tanto de investigación como para usuarios finales, desarrollando una serie de sistemas que aprovechan la melodía extraída para la búsqueda de música por semejanza (identificación de versiones y búsqueda por tarareo), la clasificación del estilo musical, la transcripción o conversión de audio a partitura, y el análisis musical con métodos computacionales. La tesis también incluye un amplio análisis comparativo del estado de la cuestión en extracción de melodía y el primer análisis crítico existente de la metodología de evaluación de algoritmos de este tipo
La indústria musical va ser una de les primeres a veure's completament reestructurada pels avenços de la tecnologia digital, i avui en dia tenim accés a milers de cançons emmagatzemades als nostres dispositius mòbils i a milions més a través de serveis en xarxa. Al tenir aquesta immensa quantitat de música al nostre abast, necessitem noves maneres de descriure, indexar, buscar i interactuar amb el contingut musical. Aquesta tesi es centra en una tecnologia que obre les portes a noves aplicacions en aquesta àrea: l'extracció automàtica de la melodia a partir d'una gravació musical polifònica. Tot i que identificar la melodia d'una peça és quelcom que els humans podem fer relativament fàcilment, fer-ho de forma automàtica presenta una alta complexitat, ja que requereix combinar coneixement de processament del senyal, acústica, aprenentatge automàtic i percepció sonora. Aquesta tasca es coneix dins de l'àmbit d'investigació com a “extracció de melodia”, i consisteix tècnicament a estimar la seqüència de altures tonals corresponents a la melodia predominant d'una peça musical a partir de l'anàlisi del senyal d'àudio. Aquesta tesi presenta un mètode innovador per a l'extracció de la melodia basat en el seguiment i caracterització de contorns tonals. Per a fer-ho, mostrem com es poden explotar les característiques de contorns combinades amb regles basades en la percepció auditiva per a identificar la melodia a partir de tot el contingut tonal d'una gravació, tant de manera heurística com a través de models apresos automàticament. A més d'això, comprovem a través d'una iniciativa internacional d'avaluació comparativa d'algoritmes que el mètode proposat obté resultats punters. De fet, obté la precisió més alta de tots els algoritmes proposats fins la data d'avui. A demés, la tesi demostra la utilitat del mètode en diverses aplicacions tant d'investigació com per a usuaris finals, desenvolupant una sèrie de sistemes que aprofiten la melodia extreta per a la cerca de música per semblança (identificació de versions i cerca per taral•larà), la classificació de l'estil musical, la transcripció o conversió d'àudio a partitura, i l'anàlisi musical amb mètodes computacionals. La tesi també inclou una àmplia anàlisi comparativa de l'estat de l'art en extracció de melodia i la primera anàlisi crítica existent de la metodologia d'avaluació d'algoritmes d'aquesta mena.
Matos, João Soares Gonçalves de. "Poderá uma Língua Natural ser Sexista? Avaliação do Impacto de Informação Gramatical e de Estereótipos de Género na Compreensão de Enunciados com Sujeitos Masculinos Genéricos em Português Europeu." Master's thesis, 2020. http://hdl.handle.net/10362/110814.
Full textOne of the most debated topics in the literature regarding the intersection of grammatical phenomena and human gender is the analysis of Masculine Generic (MG) nouns. In grammaticallygendered languages, MG nouns are linguistic items in which grammatical gender is morphosyntactically expressed, bearing [+Human], [-Feminine] and [+Plural] semantic features, despite being semantically ambiguous: they can signify both groups of men (in their specific interpretation), and mixed-gender groups (in their generic interpretation). Several recent studies (Esaulova et al., 2015; Gygax et al., 2012; Horvath et al., 2016; Stahlberg et al., 2007; Vervecken et al., 2013) have proposed that within Indo-European grammatical gender languages (German, English, French, and others), MG nouns are preferentially interpreted as specific, not as generic. However, to the best of our knowledge, no previous research has investigated whether this is the case in European Portuguese (EP). The present dissertation will directly address this shortcoming by using two experimental paradigms. Experiment I provided norms on gender stereotypes associated with EP nouns. We adapted the methodology developed by Gabriel et al. (2008) and Miserky et al. (2014). We created a list of 93 social role names which were rated for their gender stereotypicality by 342 native EP speakers. We manipulated the scale’s direction (with female versions of names being expressed on the left side vs. the right side) and the grammatical information incorporated by the evaluated nouns (feminine and masculine versions of nouns vs. MG nouns). Our results show there was an overall tendency in the stereotipicality ratings to judge the role names as being male. This tendency was less pronounced, however, when the feminine version of nouns was presented, and even less so when this version was presented on the left side of the scale. Experiment II evaluated the preferential interpretation of EP MG nouns and the influence of gender stereotypes in such interpretations. We adapted the sentence evaluation paradigm used by Gygax et al. (2008) to EP. Test items were sequences of two sentences whose syntactic subjects were referentially bound. The first subject was an MG noun; the second subject was either a conceptually masculine noun or a conceptually feminine noun (respectively “homens” or “mulheres”). Each participant had to decide whether or not the second sentence was a possible continuation for the first. We controlled the gender stereotype of the evaluated noun (which could be masculine, feminine or neuter) as well as the conceptual gender of the second sentence’s subject. The role names’ stereotypicality had no relevant influence in the interpretation of these nouns, since such interpretations were determined by their grammatical gender. Our results confirm that: (1) the preferential reading of EP MG nouns is specific, and not generic, and that (2) it is possible to reduce women’s conceptual invisibility by explicitly referring them within nominal morphology, especially if the feminine term is located on the left periphery of the syntactic coordination (for instance, “alunas e alunos”). Further, our findings highlight the importance of the socio-praxeological context of linguistic performance, and of how the structure of a natural language can influence human cognition. We discussed these insights and their sociocultural and political repercussions are discussed in this work within the frameworks of Socio-Discursive Interactionism, Neo-Whorfianism and Feminist Linguistics.
Dias, Mariana Rebelo. "Discovery of sensitive data with natural language processing." Master's thesis, 2019. http://hdl.handle.net/10071/20905.
Full textO processo de preservação de dados sensíveis está em constante crescimento e cada vez apresenta maior importância, proveniente especialmente das diretivas e leis impostas pela União Europeia. O esforço para criar sistemas automáticos é contínuo, mas o processo é realizado na maioria dos casos de forma manual ou semiautomática. Neste trabalho desenvolvemos um componente de Extração e Classificação de dados sensíveis, que processa textos não-estruturados em Português Europeu. O objetivo consistiu em criar um sistema que permite às organizações compreender os seus dados e cumprir com fins legais de conformidade e segurança. Para resolver este problema, foi estudada uma abordagem híbrida de Reconhecimento de Entidades Mencionadas para a língua Portuguesa. Esta abordagem combina técnicas baseadas em regras e léxicos, algoritmos de aprendizagem automática e redes neuronais. As primeiras abordagens baseadas em regras e léxicos, foram utilizadas apenas para um conjunto de classes especificas. Para as restantes classes de entidades foram utilizadas as ferramentas SpaCy e Stanford NLP, testados dois modelos estatísticos — Conditional Random Fields e Random Forest – e por fim testada uma abordagem baseada em redes neuronais – Bidirectional-LSTM. Ao nível das ferramentas utilizadas os melhores resultados foram conseguidos com o modelo Stanford NER (86,41%). Através dos modelos estatísticos percebemos que o Conditional Random Fields é o que consegue obter melhores resultados, com um f1-score de 65,50%. Com a última abordagem, uma rede neuronal Bi-LSTM, conseguimos resultado de f1-score de aproximadamente 83,01%. Para o treino e teste das diferentes abordagens foram utilizados os conjuntos de dados HAREM Golden Collection, SIGARRA News Corpus e DataSense NER Corpus.