Дисертації з теми "Combinational speech"

The hybrid connectionist-hidden Markov model (HMM) approach to large vocabulary continuous speech recognition has been shown to be competitive with HMM based systems. However, the recent availability of extremely large amounts of acoustic training data has highlighted a problem with the connectionist acoustic modelling paradigm. The effective use of such large amounts of data is difficult due to the computational requirements of training large connectionist models. This dissertation details research aimed at increasing the performance of connectionist acoustic models through the effective use of available training data. The methods investigated are based on ensembles of models. An ensemble is a collection of models which are combined in a manner such that the performance of the ensemble is greater than that of any of the models which form the ensemble. Most ensemble methods use a simple linear combination of the model estimates to form the ensemble estimate. A data dependent ensemble technique has been developed in which the combination of the ensemble models is dependent on the current input. The use of ensembles for speaker adaptation has been investigated, and a method based on clustering of training data has been developed and implemented. This speaker adaptation scheme does not require additional adaptation data, and can reduce the error rate of a hybrid connectionist-HMM speaker independent recognition system by up to 14.5%. In addition, clustering data allows effective use of large amounts of training data. Boosting is a method which makes selective use of training data, and produces an ensemble with each model trained on data drawn from a different distribution. Results on the optical character recognition task suggest that boosting can provide considerable gains in classification performance. The application of boosting to acoustic modelling has been investigated, and a modified boosting procedure developed. The boosting algorithms have been applied to multilayer perceptron acoustic models, and performance of the models assessed on a number of ARPA benchmark tasks. The results show that boosting consistently provides a 14-19% reduction in word error rate. The standard boosting techniques are not suitable for use with recurrent network acoustic models, and three new boosting algorithms have been developed for use with connectionist models with internal memory. These new boosting algorithms have also been evaluated on a number of ARPA benchmark tests, and have been shown to lead to a reduction in word error rate of 10-18%.

Speech is inherently bimodal, relying on cues from the acoustic and visual speech modalities for perception. The McGurk effect demonstrates that when humans are presented with conflicting acoustic and visual stimuli, the perceived sound may not exist in either modality. This effect has formed the basis for modelling the complementary nature of acoustic and visual speech by encapsulating them into the relatively new research field of audio-visual speech processing (AVSP). Traditional acoustic based speech processing systems have attained a high level of performance in recent years, but the performance of these systems is heavily dependent on a match between training and testing conditions. In the presence of mismatched conditions (eg. acoustic noise) the performance of acoustic speech processing applications can degrade markedly. AVSP aims to increase the robustness and performance of conventional speech processing applications through the integration of the acoustic and visual modalities of speech, in particular the tasks of isolated word speech and text-dependent speaker recognition. Two major problems in AVSP are addressed in this thesis, the first of which concerns the extraction of pertinent visual features for effective speech reading and visual speaker recognition. Appropriate representations of the mouth are explored for improved classification performance for speech and speaker recognition. Secondly, there is the question of how to effectively integrate the acoustic and visual speech modalities for robust and improved performance. This question is explored in-depth using hidden Markov model(HMM)classifiers. The development and investigation of integration strategies for AVSP required research into a new branch of pattern recognition known as classifier combination theory. A novel framework is presented for optimally combining classifiers so their combined performance is greater than any of those classifiers individually. The benefits of this framework are not restricted to AVSP, as they can be applied to any task where there is a need for combining independent classifiers.

Many current psycholinguistic theories view the mental lexicon as a listing of (only) unpredictable sound-meaning correspondences (primarily words and morphemes). Under this view, regular complex word forms and syntactic structures are built with rules during language production, obviating the need for storage of complex but regular linguistic material. This type of model conflicts with recent experimental evidence that suggests that lexical memory may in fact consist of a more heterogeneous set of linguistic units, including complex word forms and multi-word expressions that in theory could be constructed via rules. Storage of such material seems to be driven largely by frequency. The present research consisted of two experiments designed to investigate whether semantically transparent noun and adjective phrases are stored as single lexical units when they are very frequent. Results from the two tasks (grammaticality judgment and speech production) supported the notion that frequent word combinations can come to be stored holistically in lexical memory. It was argued that usage-based models of lexical memory (vs. dominant generative theories) best account for such data. Data from language acquisition, aphasia, and corpus studies were offered as complementary evidence in support of the more general claim that a large component of linguistic competence is knowledge of lexical co-occurrence patterns. Finally, it was hypothesized that an exemplar-based model of lexical memory best captures the range of available data.

The present study investigates children's production of gesture-speech combinations and how they relate to language acquisition. 16 children were observed at seven age points (0;9, 1;0, 1;3, 1;6, 1;9, 2;0 and 3;0), and their gesture-speech combinations were classified into the categories complementary, supplementary and discourse combinations. The results show that the production of gesture-speech combinations over time follows different patterns in children with high, average and low productive vocabulary. Furthermore, the amount of gesture-speech combinations produced at four age points predicted productive vocabulary at the age of 2;6, and the amount produced at the age of 1;6 predicted sentence complexity at 3;0. The use of complementary gesture-speech combinations at three age points was also linked to productive vocabulary. The age of onset of supplementary gesture-speech combinations predicted sentence complexity at 3;0, while the age of onset of discourse gesture-speech combinations predicted productive vocabulary at 2;6. The results support previous research suggesting that complementary and supplementary gesture-speech combinations play an important role in child language acquisition. Additionally, the results of the present study suggest that discourse gesture-speech combinations are also connected with language development.
Den här studien undersöker barns produktion av gest-ordkombinationer och hur dessa är relaterade till språkutveckling. 16 barns produktion undersöktes vid sju ålderspunkter (0;9, 1;0, 1;3, 1;6, 1;9, 2;0 och 3;0), och deras gest-ordkombinationer klassificerades i kategorierna komplementära, supplementära och diskursiva kombinationer. Resultaten visar att produktionen av gest-ordkombinationer följer olika mönster hos barn med högt, medel och lågt produktivt ordförråd. Mängden gest-ordkombinationer som producerades vid fyra ålderspunkter predicerade produktivt ordförråd vid 2;6, och mängden gest-ordkombinationer som producerades vid 1;6 predicerade meningskomplexitet vid 3;0. Användningen av komplementära gest-ordkombinationer vid tre ålderpunkter uppvisade ett samband med produktivt ordförråd vid 2;6. Tillägnandeåldern för supplementära gest-ordkombinationer predicerade meningskomplexitet vid 3;0, medan tillägnandeåldern för diskursiva gest-ordkombinationer predicerade produktivt ordförråd vid 2;6. Resultaten i den här studien ger stöd åt tidigare forskning, som har visat att komplementära och supplementära gest-ordkombinationer spelar en viktig roll i barns språkutveckling. Därutöver visar resultaten i den här studien att diskursiva gest-ordkombinationer också har ett samband med språkutveckling.
Modeling infant language acquisition from parent-child interaction (MINT)

Children with limited speech using graphic symbols for communication often express themselves predominantly through single symbols rather than symbol combinations. This study aimed to investigate the effect of an intervention strategy that was incorporated into shared storybook reading on the production of graphic symbol combinations. Three children between the ages of 7;9 (years;months) and 10;8 with limited speech and physical impairments participated in the study. A multiple probe design across behaviours (3 different types of semantic symbol combinations) was used, replicated across the 3 participants. Intervention entailed prompting the production of strategic symbol combinations (generated from a matrix) during shared storybook reading by using a prompting hierarchy. The participants’ production of combinations targeted during intervention as well as their ability to generalize to nontarget combinations from the matrix was monitored using a probe test (picture description task). All 3 participants showed some gains in acquiring the combinations and generalizing to nontarget combinations, as measured by the probe test. While 1 participant showed convincing effects, the other 2 showed lower effects. Lower effects may be partly ascribed to participant characteristics as well as to the discrepancies between the intervention and probe contexts. All participants performed better within the shared storybook reading context. Results suggest that the production of symbol combinations can be facilitated during shared storybook reading and that the matrix strategy promotes generalization to untrained semantic combinations. However, participant gains may not reflect immediately in formal testing situations.
Thesis (PhD)--University of Pretoria, 2012.
Centre for Augmentative and Alternative Communication (CAAC)
unrestricted

Au long des dernières décennies, des importants avancements ont été réalisés dans le domaine de la reconnaissance de la parole à grand vocabulaire. Un des défis à relever dans le domaine concerne la réduction des coûts de développement nécessaires pour construire un nouveau système ou adapter un système existant à une nouvelle tâche, langue ou dialecte. Les systèmes de reconnaissance de la parole à l’état de l’art sont basés sur les principes de l’apprentissage statistique, utilisant l’information fournie par deux modèles stochastiques, un modèle acoustique (MA) et un modèle de langue (ML). Les méthodes standards utilisées pour construire ces modèles s’appuient sur deux hypothèses de base : les jeux de données d’apprentissage sont suffisamment grands, et les données d’apprentissage correspondent bien à la tâche cible. Il est bien connu qu’une partie importante des coûts de développement est dû à la préparation des corpora qui remplissent ces deux conditions, l’origine principale des coûts étant la transcription manuelle des données audio. De plus, pour certaines applications, notamment la reconnaissance des langues et dialectes dits "peu dotés", la collecte des données est en soi une mission difficile. Cette thèse a pour but d’examiner et de proposer des méthodes visant à réduire le besoin de transcriptions manuelles des données audio pour une tâche donnée. Deux axes de recherche ont été suivis. Dans un premier temps, des méthodes d’apprentissage dits "non-supervisées" sont explorées. Leur point commun est l’utilisation des transcriptions audio obtenues automatiquement à l’aide d’un système de reconnaissance existant. Des méthodes non-supervisées sont explorées pour la construction de trois des principales composantes des systèmes de reconnaissance. D’abord, une nouvelle méthode d’apprentissage non-supervisée des MAs est proposée : l’utilisation de plusieurs hypothèses de décodage (au lieu de la meilleure uniquement) conduit à des gains de performance substantiels par rapport à l’approche standard. L’approche non-supervisée est également étendue à l’estimation des paramètres du réseau de neurones (RN) utilisé pour l’extraction d’attributs acoustiques. Cette approche permet la construction des modèles acoustiques d’une façon totalement non-supervisée et conduit à des résultats compétitifs en comparaison avec des RNs estimés de façon supervisée. Finalement, des méthodes non-supervisées sont explorées pour l’estimation des MLs à repli (back-off ) standards et MLs neuronaux. Il est montré que l’apprentissage non-supervisée des MLs conduit à des gains de performance additifs (bien que petits) à ceux obtenus par l’apprentissage non-supervisée des MAs. Dans un deuxième temps, cette thèse propose l’utilisation de l’interpolation de modèles comme une alternative rapide et flexible pour la construction des MAs pour une tâche cible. Les modèles obtenus à partir d’interpolation se montrent plus performants que les modèles de base, notamment ceux estimés à échantillons regroupés ou ceux adaptés à la tâche cible. Il est montré que l’interpolation de modèles est particulièrement utile pour la reconnaissance des dialectes peu dotés. Quand la quantité de données d’apprentissage acoustiques du dialecte ciblé est petite (2 à 3 heures) ou même nulle, l’interpolation des modèles conduit à des gains de performances considérables par rapport aux méthodes standards
One of the outstanding challenges in large vocabulary automatic speech recognition (ASR) is the reduction of development costs required to build a new recognition system or adapt an existing one to a new task, language or dialect. The state-of-the-art ASR systems are based on the principles of the statistical learning paradigm, using information provided by two stochastic models, an acoustic (AM) and a language (LM) model. The standard methods used to estimate the parameters of such models are founded on two main assumptions : the training data sets are large enough, and the training data match well the target task. It is well-known that a great part of system development costs is due to the construction of corpora that fulfill these requirements. In particular, manually transcribing the audio data is the most expensive and time-consuming endeavor. For some applications, such as the recognition of low resourced languages or dialects, finding and collecting data is also a hard (and expensive) task. As a means to lower the cost required for ASR system development, this thesis proposes and studies methods that aim to alleviate the need for manually transcribing audio data for a given target task. Two axes of research are explored. First, unsupervised training methods are explored in order to build three of the main components of ASR systems : the acoustic model, the multi-layer perceptron (MLP) used to extract acoustic features and the language model. The unsupervised training methods aim to estimate the model parameters using a large amount of automatically (and inaccurately) transcribed audio data, obtained thanks to an existing recognition system. A novel method for unsupervised AM training that copes well with the automatic audio transcripts is proposed : the use of multiple recognition hypotheses (rather than the best one) leads to consistent gains in performance over the standard approach. Unsupervised MLP training is proposed as an alternative to build efficient acoustic models in a fully unsupervised way. Compared to cross-lingual MLPs trained in a supervised manner, the unsupervised MLP leads to competitive performance levels even if trained on only about half of the data amount. Unsupervised LM training approaches are proposed to estimate standard back-off n-gram and neural network language models. It is shown that unsupervised LM training leads to additive gains in performance on top of unsupervised AM training. Second, this thesis proposes the use of model interpolation as a rapid and flexible way to build task specific acoustic models. In reported experiments, models obtained via interpolation outperform the baseline pooled models and equivalent maximum a posteriori (MAP) adapted models. Interpolation proves to be especially useful for low resourced dialect ASR. When only a few (2 to 3 hours) or no acoustic data truly matching the target dialect are available for AM training, model interpolation leads to substantial performance gains compared to the standard training methods

Dissertation (Ph.D. in Aeronautical and Astronautical Engineering)--Naval Postgraduate School, Sept. 2001.
Dissertation supervisor: Platzer, Max F. "September 2001." Includes bibliographical references (p. 171-174). Also available in print.

Thesis (Ph. D.)--Worcester Polytechnic Institute.
Keywords: design guidelines; NCHRP Report 350; finite element analysis; impact; curbs; guardrail; curb and barrier combinations. Includes bibliographical references (p. 310-316).

The objective of the present study was to establish effective combinations of marching speed and backpack load in order to meet specific military requirements. Thirty infantrymen from the South African National Defence Force (SANDF) comprised the sample and experimental procedures were conducted in a laboratory setting using a Cybex Trotter treadmill. Sixteen conditions were set up which included combinations of four speeds (3.5, 4.5, 5.5, and 6.5 km.h⁻¹) and four backpack loads (20, 35, 50, and 65kg). Each subject was required to complete 8 of the sixteen conditions, each consisting of a six-minute treadmill march. Physiological data (heart rate, ventilation and metabolic responses), kinematic gait responses (step-rate and stride length) and perceptions of exertion (“Central” and “Local” RPE) were collected during the third and sixth minutes of the treadmill march and areas of body discomfort were identified post-march. Responses revealed five distinct categories of exertional strain. Three marches constituted “nominal” (below 40% VO₂max) and three “excessive” strain (above 75% VO₂ max). These represent combinations of extreme military demands and are highly unlikely to be utilised by the military. Three “tolerable” levels of required effort were recommended and these 10 combinations were further divided into three sub-categories. The “moderate” stress marches were identified as “ideal” for prolonged marches and had statistically similar responses of working heart rates (range of 118 bt.min⁻¹ to 127 bt.min⁻¹), energy expenditure (26 kJ.min⁻¹ and 27 kJ.min⁻¹) and ratings of perceived exertion (“Central” ratings of 10 and 11). Thus, marching at 5.5 km.h⁻¹with 20kg, 4.5 km.h⁻¹ with 35kg or 3.5 km.h⁻¹ with 50kg all require a similar energy cost. Four “heavy” category marches were identified for possible use when the duration of the march is reduced. During these marches responses were statistically similar with heart rates ranging from 127 bt.min⁻¹ to 137 bt.min⁻¹, energy expenditure from 32 kJ.min⁻¹ to 37 kJ.min⁻¹ and “Central” ratings of perceived exertion were 12 and 13. When short, high intensity marches are necessary, then combinations from the “very heavy” category may be utilised but with caution. During these marches, soldiers were taxed between 65% and 75% of VO2 max. The results of this study clearly demonstrate that the interplay between speed and load needs to be adjusted when determining “ideal” combinations for specific military demands. Essentially, if speed is of the essence then load must be reduced, and if heavy loads need to be transported then speed must be reduced.

Natural Language Processing (NLP) is an interdisciplinary research field of Computer Science, Linguistics, and Pattern Recognition that studies, among others, the use of human natural languages in Human-Computer Interaction (HCI). Most of NLP research tasks can be applied for solving real-world problems. This is the case of natural language recognition and natural language translation, that can be used for building automatic systems for document transcription and document translation. Regarding digitalised handwritten text documents, transcription is used to obtain an easy digital access to the contents, since simple image digitalisation only provides, in most cases, search by image and not by linguistic contents (keywords, expressions, syntactic or semantic categories). Transcription is even more important in historical manuscripts, since most of these documents are unique and the preservation of their contents is crucial for cultural and historical reasons. The transcription of historical manuscripts is usually done by paleographers, who are experts on ancient script and vocabulary. Recently, Handwritten Text Recognition (HTR) has become a common tool for assisting paleographers in their task, by providing a draft transcription that they may amend with more or less sophisticated methods. This draft transcription is useful when it presents an error rate low enough to make the amending process more comfortable than a complete transcription from scratch. Thus, obtaining a draft transcription with an acceptable low error rate is crucial to have this NLP technology incorporated into the transcription process. The work described in this thesis is focused on the improvement of the draft transcription offered by an HTR system, with the aim of reducing the effort made by paleographers for obtaining the actual transcription on digitalised historical manuscripts. This problem is faced from three different, but complementary, scenarios: · Multimodality: The use of HTR systems allow paleographers to speed up the manual transcription process, since they are able to correct on a draft transcription. Another alternative is to obtain the draft transcription by dictating the contents to an Automatic Speech Recognition (ASR) system. When both sources (image and speech) are available, a multimodal combination is possible and an iterative process can be used in order to refine the final hypothesis. · Interactivity: The use of assistive technologies in the transcription process allows one to reduce the time and human effort required for obtaining the actual transcription, given that the assistive system and the palaeographer cooperate to generate a perfect transcription. Multimodal feedback can be used to provide the assistive system with additional sources of information by using signals that represent the whole same sequence of words to transcribe (e.g. a text image, and the speech of the dictation of the contents of this text image), or that represent just a word or character to correct (e.g. an on-line handwritten word). · Crowdsourcing: Open distributed collaboration emerges as a powerful tool for massive transcription at a relatively low cost, since the paleographer supervision effort may be dramatically reduced. Multimodal combination allows one to use the speech dictation of handwritten text lines in a multimodal crowdsourcing platform, where collaborators may provide their speech by using their own mobile device instead of using desktop or laptop computers, which makes it possible to recruit more collaborators.
El Procesamiento del Lenguaje Natural (PLN) es un campo de investigación interdisciplinar de las Ciencias de la Computación, Lingüística y Reconocimiento de Patrones que estudia, entre otros, el uso del lenguaje natural humano en la interacción Hombre-Máquina. La mayoría de las tareas de investigación del PLN se pueden aplicar para resolver problemas del mundo real. Este es el caso del reconocimiento y la traducción del lenguaje natural, que se pueden utilizar para construir sistemas automáticos para la transcripción y traducción de documentos. En cuanto a los documentos manuscritos digitalizados, la transcripción se utiliza para facilitar el acceso digital a los contenidos, ya que la simple digitalización de imágenes sólo proporciona, en la mayoría de los casos, la búsqueda por imagen y no por contenidos lingüísticos. La transcripción es aún más importante en el caso de los manuscritos históricos, ya que la mayoría de estos documentos son únicos y la preservación de su contenido es crucial por razones culturales e históricas. La transcripción de manuscritos históricos suele ser realizada por paleógrafos, que son personas expertas en escritura y vocabulario antiguos. Recientemente, los sistemas de Reconocimiento de Escritura (RES) se han convertido en una herramienta común para ayudar a los paleógrafos en su tarea, la cual proporciona un borrador de la transcripción que los paleógrafos pueden corregir con métodos más o menos sofisticados. Este borrador de transcripción es útil cuando presenta una tasa de error suficientemente reducida para que el proceso de corrección sea más cómodo que una completa transcripción desde cero. Por lo tanto, la obtención de un borrador de transcripción con una baja tasa de error es crucial para que esta tecnología de PLN sea incorporada en el proceso de transcripción. El trabajo descrito en esta tesis se centra en la mejora del borrador de transcripción ofrecido por un sistema RES, con el objetivo de reducir el esfuerzo realizado por los paleógrafos para obtener la transcripción de manuscritos históricos digitalizados. Este problema se enfrenta a partir de tres escenarios diferentes, pero complementarios: · Multimodalidad: El uso de sistemas RES permite a los paleógrafos acelerar el proceso de transcripción manual, ya que son capaces de corregir en un borrador de la transcripción. Otra alternativa es obtener el borrador de la transcripción dictando el contenido a un sistema de Reconocimiento Automático de Habla. Cuando ambas fuentes están disponibles, una combinación multimodal de las mismas es posible y se puede realizar un proceso iterativo para refinar la hipótesis final. · Interactividad: El uso de tecnologías asistenciales en el proceso de transcripción permite reducir el tiempo y el esfuerzo humano requeridos para obtener la transcripción correcta, gracias a la cooperación entre el sistema asistencial y el paleógrafo para obtener la transcripción perfecta. La realimentación multimodal se puede utilizar en el sistema asistencial para proporcionar otras fuentes de información adicionales con señales que representen la misma secuencia de palabras a transcribir (por ejemplo, una imagen de texto, o la señal de habla del dictado del contenido de dicha imagen de texto), o señales que representen sólo una palabra o carácter a corregir (por ejemplo, una palabra manuscrita mediante una pantalla táctil). · Crowdsourcing: La colaboración distribuida y abierta surge como una poderosa herramienta para la transcripción masiva a un costo relativamente bajo, ya que el esfuerzo de supervisión de los paleógrafos puede ser drásticamente reducido. La combinación multimodal permite utilizar el dictado del contenido de líneas de texto manuscrito en una plataforma de crowdsourcing multimodal, donde los colaboradores pueden proporcionar las muestras de habla utilizando su propio dispositivo móvil en lugar de usar ordenadores,
El Processament del Llenguatge Natural (PLN) és un camp de recerca interdisciplinar de les Ciències de la Computació, la Lingüística i el Reconeixement de Patrons que estudia, entre d'altres, l'ús del llenguatge natural humà en la interacció Home-Màquina. La majoria de les tasques de recerca del PLN es poden aplicar per resoldre problemes del món real. Aquest és el cas del reconeixement i la traducció del llenguatge natural, que es poden utilitzar per construir sistemes automàtics per a la transcripció i traducció de documents. Quant als documents manuscrits digitalitzats, la transcripció s'utilitza per facilitar l'accés digital als continguts, ja que la simple digitalització d'imatges només proporciona, en la majoria dels casos, la cerca per imatge i no per continguts lingüístics (paraules clau, expressions, categories sintàctiques o semàntiques). La transcripció és encara més important en el cas dels manuscrits històrics, ja que la majoria d'aquests documents són únics i la preservació del seu contingut és crucial per raons culturals i històriques. La transcripció de manuscrits històrics sol ser realitzada per paleògrafs, els quals són persones expertes en escriptura i vocabulari antics. Recentment, els sistemes de Reconeixement d'Escriptura (RES) s'han convertit en una eina comuna per ajudar els paleògrafs en la seua tasca, la qual proporciona un esborrany de la transcripció que els paleògrafs poden esmenar amb mètodes més o menys sofisticats. Aquest esborrany de transcripció és útil quan presenta una taxa d'error prou reduïda perquè el procés de correcció siga més còmode que una completa transcripció des de zero. Per tant, l'obtenció d'un esborrany de transcripció amb un baixa taxa d'error és crucial perquè aquesta tecnologia del PLN siga incorporada en el procés de transcripció. El treball descrit en aquesta tesi se centra en la millora de l'esborrany de la transcripció ofert per un sistema RES, amb l'objectiu de reduir l'esforç realitzat pels paleògrafs per obtenir la transcripció de manuscrits històrics digitalitzats. Aquest problema s'enfronta a partir de tres escenaris diferents, però complementaris: · Multimodalitat: L'ús de sistemes RES permet als paleògrafs accelerar el procés de transcripció manual, ja que són capaços de corregir un esborrany de la transcripció. Una altra alternativa és obtenir l'esborrany de la transcripció dictant el contingut a un sistema de Reconeixement Automàtic de la Parla. Quan les dues fonts (imatge i parla) estan disponibles, una combinació multimodal és possible i es pot realitzar un procés iteratiu per refinar la hipòtesi final. · Interactivitat: L'ús de tecnologies assistencials en el procés de transcripció permet reduir el temps i l'esforç humà requerits per obtenir la transcripció real, gràcies a la cooperació entre el sistema assistencial i el paleògraf per obtenir la transcripció perfecta. La realimentació multimodal es pot utilitzar en el sistema assistencial per proporcionar fonts d'informació addicionals amb senyals que representen la mateixa seqüencia de paraules a transcriure (per exemple, una imatge de text, o el senyal de parla del dictat del contingut d'aquesta imatge de text), o senyals que representen només una paraula o caràcter a corregir (per exemple, una paraula manuscrita mitjançant una pantalla tàctil). · Crowdsourcing: La col·laboració distribuïda i oberta sorgeix com una poderosa eina per a la transcripció massiva a un cost relativament baix, ja que l'esforç de supervisió dels paleògrafs pot ser reduït dràsticament. La combinació multimodal permet utilitzar el dictat del contingut de línies de text manuscrit en una plataforma de crowdsourcing multimodal, on els col·laboradors poden proporcionar les mostres de parla utilitzant el seu propi dispositiu mòbil en lloc d'utilitzar ordinadors d'escriptori o portàtils, la qual cosa permet ampliar el nombr
Granell Romero, E. (2017). Advances on the Transcription of Historical Manuscripts based on Multimodality, Interactivity and Crowdsourcing [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/86137
TESIS

Nous abordons, dans cette thèse, les méthodes de combinaison de systèmesde transcription de la parole à Large Vocabulaire. Notre étude se concentre surl'attelage de systèmes de transcription hétérogènes dans l'objectif d'améliorerla qualité de la transcription à latence contrainte. Les systèmes statistiquessont affectés par les nombreuses variabilités qui caractérisent le signal dela parole. Un seul système n'est généralement pas capable de modéliserl'ensemble de ces variabilités. La combinaison de différents systèmes detranscription repose sur l'idée d'exploiter les points forts de chacun pourobtenir une transcription finale améliorée. Les méthodes de combinaisonproposées dans la littérature sont majoritairement appliquées a posteriori,dans une architecture de transcription multi-passes. Cela nécessite un tempsde latence considérable induit par le temps d'attente requis avant l'applicationde la combinaison.Récemment, une méthode de combinaison intégrée a été proposée. Cetteméthode est basée sur le paradigme de décodage guidé (DDA :Driven DecodingAlgorithm) qui permet de combiner différents systèmes durant le décodage. Laméthode consiste à intégrer des informations en provenance de plusieurs systèmes dits auxiliaires dans le processus de décodage d'un système dit primaire.Notre contribution dans le cadre de cette thèse porte sur un double aspect : d'une part, nous proposons une étude sur la robustesse de la combinaison par décodage guidé. Nous proposons ensuite, une amélioration efficacement généralisable basée sur le décodage guidé par sac de n-grammes,appelé BONG. D'autre part, nous proposons un cadre permettant l'attelagede plusieurs systèmes mono-passe pour la construction collaborative, à latenceréduite, de la sortie de l'hypothèse de reconnaissance finale. Nous présentonsdifférents modèles théoriques de l'architecture d'attelage et nous exposons unexemple d'implémentation en utilisant une architecture client/serveur distribuée. Après la définition de l'architecture de collaboration, nous nous focalisons sur les méthodes de combinaison adaptées à la transcription automatiqueà latence réduite. Nous proposons une adaptation de la combinaison BONGpermettant la collaboration, à latence réduite, de plusieurs systèmes mono-passe fonctionnant en parallèle. Nous présentons également, une adaptationde la combinaison ROVER applicable durant le processus de décodage via unprocessus d'alignement local suivi par un processus de vote basé sur la fréquence d'apparition des mots. Les deux méthodes de combinaison proposéespermettent la réduction de la latence de la combinaison de plusieurs systèmesmono-passe avec un gain significatif du WER.

La séparation aveugle de sources audio dans le cas sous-déterminé est un problème mathématique complexe dont il est aujourd'hui possible d'obtenir une solution satisfaisante, à condition de sélectionner la méthode la plus adaptée au problème posé et de savoir paramétrer celle-ci soigneusement. Afin d'automatiser cette étape de sélection déterminante, nous proposons dans cette thèse de recourir au principe de fusion. L'idée est simple : il s'agit, pour un problème donné, de sélectionner plusieurs méthodes de résolution plutôt qu'une seule et de les combiner afin d'en améliorer la solution. Pour cela, nous introduisons un cadre général de fusion qui consiste à formuler l'estimée d'une source comme la combinaison de plusieurs estimées de cette même source données par différents algorithmes de séparation, chaque estimée étant pondérée par un coefficient de fusion. Ces coefficients peuvent notamment être appris sur un ensemble d'apprentissage représentatif du problème posé par minimisation d'une fonction de coût liée à l'objectif de séparation. Pour aller plus loin, nous proposons également deux approches permettant d'adapter les coefficients de fusion au signal à séparer. La première formule la fusion dans un cadre bayésien, à la manière du moyennage bayésien de modèles. La deuxième exploite les réseaux de neurones profonds afin de déterminer des coefficients de fusion variant en temps. Toutes ces approches ont été évaluées sur deux corpus distincts : l'un dédié au rehaussement de la parole, l'autre dédié à l'extraction de voix chantée. Quelle que soit l'approche considérée, nos résultats montrent l'intérêt systématique de la fusion par rapport à la simple sélection, la fusion adaptative par réseau de neurones se révélant être la plus performante
Underdetermined blind source separation is a complex mathematical problem that can be satisfyingly resolved for some practical applications, providing that the right separation method has been selected and carefully tuned. In order to automate this selection process, we propose in this thesis to resort to the principle of fusion which has been widely used in the related field of classification yet is still marginally exploited in source separation. Fusion consists in combining several methods to solve a given problem instead of selecting a unique one. To do so, we introduce a general fusion framework in which a source estimate is expressed as a linear combination of estimates of this same source given by different separation algorithms, each source estimate being weighted by a fusion coefficient. For a given task, fusion coefficients can then be learned on a representative training dataset by minimizing a cost function related to the separation objective. To go further, we also propose two ways to adapt the fusion coefficients to the mixture to be separated. The first one expresses the fusion of several non-negative matrix factorization (NMF) models in a Bayesian fashion similar to Bayesian model averaging. The second one aims at learning time-varying fusion coefficients thanks to deep neural networks. All proposed methods have been evaluated on two distinct corpora. The first one is dedicated to speech enhancement while the other deals with singing voice extraction. Experimental results show that fusion always outperform simple selection in all considered cases, best results being obtained by adaptive time-varying fusion with neural networks

Telemetry system, Optimisation, Sensoric networks, Smart Grid, Internet of Things, Sensors, Information security, Cryptography, Cryptography algorithms, Cryptosystem, Confidentiality, Integrity, Authentication, Data freshness, Non-Repudiation.

Family environment plays a crucial role in bilingual language socialization in early childhood. The bilingual family introduces the child not only to the languages-in-acquisition, but also to the preferred language use patterns. In recent years the discussion on how and when a bilingual child comes to use her two languages in contextually appropriate ways has become central to Bilingual First Language Acquisition (BFLA). While evidence for language differentiation and sensitivity to interlocutor’s preferred language is available in the two-word stage, few studies consider the language learning environment and its impact on bilingual development in the one-word stage and early combinatorial speech. This longitudinal case study reports on linguistic developments from birth to 2;0 in a child who was acquiring Slovak and English simultaneously in the home. The effects of the child’s language learning environment on linguistic development from the onset of speech were considered, focusing on lexical development, word combinations, emerging morpho-syntax, and pragmatic aspects such as language choice and mixing. The child’s two languages developed separately in a side-by-side fashion, as shown by use of translation equivalents and language specific morphological markers from the beginning. She used the two languages in contextually sensitive ways from the one-word stage, relying on several pragmatic language choice strategies. Mixing was productive and accounted only for a small proportion of productions. It was explained by sociolinguistic as well as psycholinguistic factors. Language differentiation thus emerged as grammatical as well as pragmatic differentiation at the end of the one word stage.
Thesis (Ph.D.) -- University of Adelaide, School of Humanities, 2013.

碩士
國立臺北科技大學
電腦與通訊研究所
99
The performance of an automatic speech processing system is often degraded due to the embedded noise in the processed speech signal. Therefore, the speech enhancement technology is applied to the automatic speech processing systems to reduce noise interference and increase system efficiency. In this thesis, we propose a speech enhancement system that reduces the background noise by combining both the spectral magnitude estimators for speech and noise. The a priori signal-to-noise ratio (SNR) is refined by two-step noise reduction (TSNR) to remove the drawbacks of the decision-directed approach. However, there still remains residual noise in the enhanced speech. In order to solve this problem, we add a postfilter in the terminal of the system to eliminate residual noise for speech pauses. Finally, we take two objective measures (the segmental SNR and the perceptual evaluation of speech quality, PESQ) to assess the quality of the enhanced speech. Experimental results show the effectiveness of the proposed speech enhancement system.

碩士
國立成功大學
電機工程研究所
83
Most of the researches on speech signal coding have put emphasis on the speech quality and the character of source coding , but seldom taked into consideration the effect of channel noise to speech signal transmission. Due to this reason , we combined source coding and channel coded modulation to obtain the better transmission performance without expanding bandwidth. We have studied several broadly accepted speech standards-- ADPCM, CELP,and American IS-96 (speech coding standard for IS -95), and learned the source coding principles and procedures. We have also studied the schemes to estimate the speech quality and selected for the proper methods that can be used in our works. We then used the simulations to find the impacts of individual source coding bit on the speech quality by controlling error bits and divide the coding bits into several groups according to their significance on the quality. The UEP coded modulation scheme is then adopted to give unequal protection capability for different bit groups. Combining such source coding and UEP coded modulation, we can achieve an improvement of communication quality.

碩士
國立中央大學
電機工程學系
102
This thesis deals with the improvement on the speech feature extracting part in speech recognition. Feature extraction is a very important part in speech recognition, by having two advantages of reducing the amount of data and highlighting the characteristics of voice. Many researchers have been published different extracting methods or improving methods for speech features for highlighting different characteristics of voice. This thesis presents a method for combining different speech features, and binding the characteristics of different feature methods together. The result of our experiments showed that the proposed method improves the recognition rate of the keyword spotting system, and also proved that the method can effectively improve the characteristics of voice. In the second part of this thesis, we apply the keyword spotting system to iPhone smartphone app and build a voice-controlled game to achieve real-time speech recognition.

The growing need for designing and implementing reliable voice-based human-machine interfaces has inspired intensive research work in the field of voice-enabled systems, and greater robustness and reliability are being sought for those systems. Speech recognition has become ubiquitous. Automated call centers, smart phones, dictation and transcription software are among the many systems currently being designed and involving speech recognition. The need for highly accurate and optimized recognizers has never been more crucial. The research community is very actively involved in developing powerful techniques to combine the existing feature extraction methods for a better and more reliable information capture from the analog signal, as well as enhancing the language and acoustic modeling procedures to better adapt for unseen or distorted speech signal patterns. Most researchers agree that one of the most promising approaches for the problem of reducing the Word Error Rate (WER) in large vocabulary speech transcription, is to combine two or more speech recognizers and then generate a new output, in the expectation that it provides a lower error rate. The research work proposed here aims at enhancing and boosting even further the performance of the well-known Recognizer Output Voting Error Reduction (ROVER) combination technique. This is done through its integration with an error filtering approach. The proposed system is referred to as cROVER, for context-augmented ROVER. The principal idea is to flag erroneous words following the combination of the word transition networks through a scanning process at each slot of the resulting network. This step aims at eliminating some transcription errors and thus facilitating the voting process within ROVER. The error detection technique consists of spotting semantic outliers in a given decoder's transcription output. Due to the fact that most error detection techniques suffer from a high false positive rate, we propose to combine the error filtering techniques to compensate for the poor performance of each of the individual error classifiers. Experimental results, have shown that the proposed cROVER approach is able to reduce the relative WER by almost 10% through adequate combination of speech decoders. The approaches proposed here are generic enough to be used by any number of speech decoders and with any type of error filtering technique. A novel voting mechanism has also been proposed. The new confidence-based voting scheme has been inspired from the cROVER approach. The main idea consists of using the confidence scores collected from the contextual analysis, during the scoring of each word in the transition network. The new voting scheme outperformed ROVER's original voting, by up to 16% in terms of relative WER reduction.

碩士
國立臺灣大學
電信工程學研究所
94
Substantial efforts have been made in various areas towards the goal of improving the performance of large vocabulary continuous speech recognition (LVCSR) technologies. Two important areas towards this goal, among many others, are rescoring over the word graph as well as combination of multiple systems. In this thesis, we focused on these two areas for complete studies. In the area of rescoring by discriminative decoding, we studied Minimum Bayes Risk decoding (MBR), Segment Minimum Bayes Risk decoding (SMBR) [16] , Minimum Time Frame Error decoding[17], and Optimal Bayes Classification decoding (OBC)[18] with experiments on Chinese broadcast news corpus. For combination of the outputs of several different systems, we focused on the ROVER technique with N-Best input[9][20]. A new concept of integrated hypothesis space for large vocabulary continuous speech recognition (LVCSR) systems combination is then proposed. Unlike the conventional systems combination approaches such as ROVER, the hypothesis spaces are directly integrated here without string alignment. In this way the timing information for all word hypotheses is well preserved and the new framework is more flexible on rescoring approaches used. Four different rescoring criteria on the integrated hypothesis space were further explored and experiments on Chinese broadcast news corpus indicated improved performance.

碩士
國立成功大學
製造資訊與系統研究所
102
The main purpose of this study is to develop a technology for a Part-of-Speech (POS) combination-aided Chinese electronic Word-of-Mouth (eWOM) analysis. Using this technology to analyze a Chinese eWOM, we obtain eWOM sentences of the target product, and analyze eWOM polarities supplemented by POS combination rules, and constantly update analysis criteria to keep up with new network articles. Hope this technology can effectively help companies quickly and correctly understand the evaluation situation of the current eWOM, and further form a company basis for improving their relations with customers. This study mainly include: (1) Exploration of Research Domains and Related Technologies, (2) POS Combination Rules Generation and Topic Terms Extraction Method Design, (3) eWOM Identification Method Design, and (4) Implementation of POS Combination-Aided Chinese eWOM Analysis Mechanisms.

碩士
大同大學
資訊工程學系(所)
100
The speech emotion recognition is one of the important researches on the discovery of human emotion. In this thesis, we proposed a method which fuses the results of the keyword analysis and speech emotion recognition to improve the recognition rate. The speech emotion recognition model extracts features of speech signal and uses a proper classifier to classify the emotion. In this research, we use the Mel-Frequency Cepstral Coefficients (MFCC) extracted from the input speech as the feature for classification. The weighted discrete K-Nearest Neighbor (WD-KNN) and Support Vector Machine (SVM) classifiers are adopted to classify the emotion in the speech. In emotion keyword analysis, the emotion keywords were selected from the scripts. The emotion keyword category and intensity are manually defined by annotators. Finally, we fuse the results of the speech emotion recognition model and the emotion keyword analysis to improve the recognition rate. Experimental results show that the recognition rate increases 3% with the fusing model as compared to the use of the speech emotion recognition along. The incorrect rate of misclassifying non-anger to anger is reduced by 40% .

碩士
國立中興大學
電機工程學系所
94
Birdsong and Human speech recognition by Recurrent Fuzzy Network Filtered Hidden Markov Model (RFNF-HMM) in variable noise-level environments is proposed in this thesis. Birdsong and Human Speech are generated in a similar way and a similar recognition approach is proposed in this thesis. For birdsong recognition, the energy parameter is used to segment a significant portion from a birdsong sequence. Then, the linear predictive coding (LPC) coefficients of each frame in the segmented birdsong are extracted and used as feature vectors. These feature vectors are fed as inputs to HMM recognizers. Birdsong in outside practical environments are usually corrupted by non-stationary noise or other interference, which degrades recognition performance. To handle this problem, RFNF-HMM recognizer is proposed, where each HMM is connected with a RFNF for noise filtering in feature domain. Experiments in recognition of ten species of birds in Taiwan according to birdsongs are performed. For Human speech recognition in variable noise-level environments, Wavelet Energy (WE) and Zero Crossing Rate (ZCR) are proposed as detection parameters for word boundary detection. Three kinds of Intelligent Learning Networks (ILN) are used as detectors with their performance being compared. They are RSONFIN, Fuzzy System learned through the combination of Fuzzy Clustering and Support Vector Machine (FS-FCSVM), and Gaussian kernel SVM. After words detection, the cepstral coefficients are used as features and RFNF-HMM are used as recognizers. In experiments, words detection performance is evaluated both by Receiver Operator Characteristic (ROC) curves and recognition rates. Experimental results show that the proposed detection parameters are robust and effective with comparisons to the Refined Time-Frequency (RTF) parameters. Admiring recognition rates in human speech recognition with different types of unknown noises in variable noise-level environments are achieved in experiments.

This study examined the metabolic cost of South African soldiers carrying loads under various conditions of gradient and marching speed. Thirty-two male soldiers participated in the study. Three speed and load combinations, 4, 5 and 6 km.h[superscript (-1)] carrying of 50, 35 and 20 kg respectively, were imposed, on each of three gradients: -10%, 0% and +10%; a total of nine experimental conditions. Subjects wore standardized military uniforms and breathed into a portable ergospirometer (Metamax) for the duration of each condition. Subjects were required to march for six minutes under each condition, and metabolic responses were monitored during the third and sixth minute of each condition. The metabolic responses to the three level marching conditions showed no significant differences. Marching uphill resulted in a significant increase in metabolic demands under all three speed and load onditions. Downhill marching elicited significant decreases only under the two lighter load conditions. Downhill marching with heavy loads appears to show no reduction in metabolic demands placed on soldiers.

Thirty seasoned foot-soldiers were exposed to a total of 16 combinations of marching speed and backpack load, each soldier experiencing 8 of the combinations between 3.5km.h-1 with a 20kg load and 6.5km.h-1 with a 65kg load. Detailed physiological analyses were conducted, of which the heart rate responses are reported here. Steady state was achieved with all combinations under 5.5km.h-1 speed and 35kg load. Excessively elevated heart rates were elicited by all loads over 20kg at 6.5km.h-1 marching speed; by loads over 35kg at 5.5km.h-1 marching speed and by the 65kg load even at 4.5km.h-1 marching speed.

碩士
國立成功大學
交通管理科學系
102
The purpose of this paper is to investigate the traveler’s choice behavior towards two combined perishable products services-evidence from high-speed rail (HSR) and car rental industry services. This study adopts the stated preference method to obtain the possible alternatives with combinations of attribute variable levels. According to characteristics of travelers, in order to understand choice behavior of different types of travelers, the objects of this study will be divided into business travelers and leisure tourists. This study adopts the hybrid discrete choice model (HDCM) which includes psychological factors, to investigate the impact of mode choice preference under different types of travelers. Our results indicates that the higher degree in the perceived costs of business travelers are less preferred HSR and car rental combination as well as driving by self; however, the higher degree in the perceived costs of leisure tourists prefer HSR and car rental combination as well as driving by self. The higher degree in the involvement and information source of business travelers prefer HSR and car rental combination as well as driving by self. The higher degree in the perceived value of leisure tourists are less preferred HSR and car rental combination as well as driving by self.

碩士
國立體育學院
教練研究所
87
The purpose of this study was to probe into the Chen''s Passive Repeatedly Plyometric Power Machine which was designed by Chuan-Show Chen and Tyzz-Yuang Shiang. Seventeen baseball male players were selected as the subjects, and divided into three groups: Experimental groupⅠ: control weight and increase speed (6 people). Experimental groupⅡ: control speed and increase weight (6 people). Comparison group: traditional composite training (5 people). Experimental groups were trained by Chen''s Passive Repeatedly Plyometric Power Machine, and the Smith Leg Press weight training was used for Comparison group. Three times a week over six weeks lower extremities strength training was applied for all subjects. All subjects were tested of maximal strength of lower extremity, power, body shape. The results of the study are noted as followed: 1. Maximal strength of lower extremity: There were significant increase in Experimental groups and comparison group (P<0.05), and but there wasn''t significant difference between three groups. 2.Power: There was significant increase in vertical-jump, rebound-jump and standing jump of Experimental groupⅠ(P<0.05). There were significant increase in vertical-jump and rebound-jump of Experimental groupⅡ (P<0.05). There were significant increase in standing jump of Comparison group (P<0.05), but there wasn''t significant difference between three groups. 3.Body shape: There wasn''t significant increase in gluteus and thigh length of Experimental groupⅠ. There was significant increase in gluteus length of Experimental group (P<0.05). There was significant increase in left thigh, left leg, and right leg length (P<0.05) of Comparison group, but there wasn''t significant difference between three groups. According to results, we concluded as followed: 1. Chen''s Passive Repeatedly Plyometric Power Machine can improve power. 2.There is much more training effect of power in Experimental groupⅠthan Experimental groupⅡ. 3.There is much more training effect of strength in Experimental groupⅡthan Experimental groupⅠ. 4.Comparison group may have better training effect in isotonic strength training. 5.Chen''s Passive Repeatedly Plyometric Power Machine may have better training effect in gluteus than traditional composite training.

碩士
國立體育大學
運動技術研究所
100
Purpose: To investigate the effect different combination of training 100m athletes for maximum speed and stability and aerobic glycolysis results. Methods: The subjects were five boys and three girls track athletes, the average age was 13.51 ± 0.59 yr; height and weight were 163.25 ± 7.92cm, 48.5 ± 7.98kg; years of training 3.12 ± 1.1yr. Training time for two weeks, high-intensity training four times a week, each training curriculum (30m-80m-30m-80m, 80m-30m-80m-30m, 30m-30m-80m-80m, 30m-Rep.3min 80m-Rep. 5min, the Set 10min) per trip intermittent recorded the heart rate of players. All subjects proceed to the basic endurance (2-4 mmol/l) test and special test 100m section speed before the experiment and after, lactic acid and heart rate were recorded at the end of the experiment. The information obtained by SPSS for Windows 15.0 statistical software and curve analysis of SigmaPlot 8.0 software production and paired t-test this study is a significant test as p>0.05. The following conclusions: (1) After two weeks at different distances combination of training 100m measured progress than the pre-test speed of 0.46 m/s, the maximum lactic acid accumulation in the T2 extended to the first 10min. (2) After two weeks at different distances combination of training, the fastest speed show up at 40m of 100m sectional speed, till 100m decline to-0.51m/s, progress 0.12m/s than the pre-test. of -0.63.

碩士
逢甲大學
電機工程學系
107
In recent years, as a result of Taiwan's energy transformation, it is expected that the share of renewable energy generation can be increased, and wind power generation is one of them. Taiwan is now actively promoting the establishment of offshore wind farms, but the impact of wind power generation, in addition to the wind turbine itself unit equipment, topography, weather factors, wind speed is also a major factor affecting wind power generation. With the progress and development of artificial intelligence, this paper uses deep learning to make predictions. Deep learning includes a variety of models, such as multi-layer perceptrons, convolutional neural networks, and recursive neural networks. Different neural networks are used to combine different models to analyze individual predictionresults and optimize their models, and select the best models among the trained models. The paper will input different weather data and use three methods to predict wind speed.The first method is convolutional neural network with long short-term memory, the second method is convolutional neural network with bidirectional long short-term memory and the third method is convolutional neural network with bidirectional long short-term memory and principal component analysis.Using the ability of convolutional neural network to extract images.Long short-term memory and bidirectional long short-term memory are time series neural networks, Using of the principal component analysis extracted weather data feature value modeling helps to reduce computation and storage capacity, and reduce the complexity of the model, thereby improving the widespread use of the model capacity.

Dissertation submitted in partial compliance with the requirements for the Master's Degree in Technology: Chiropractic, Durban University of Technology, 2008.
To compare trunk flexion and lateral flexion range of motion (ROM) pre-post intervention. To compare the bowling speed of Action Cricket fast bowlers pre-post intervention. To determine the correlation between change in bowling speed immediately post-intervention to change in trunk flexion and lateral flexion ROM immediately post-intervention. To determine the association between change in bowling speed immediately post-intervention and the subjects’ perception of change in bowling speed. Methods: Forty asymptomatic Action Cricket fast bowlers were divided into four groups of ten each. Group 1 received lumbar spine manipulation, Group 2 received thoracic spine manipulation, Group 3 received combined thoracic and lumbar spine manipulation and Group 4 received the sham laser intervention (placebo). Pre- and post-intervention trunk flexion and lateral flexion ROM and bowling speed were measured using a digital inclinometer and a SpeedTracTM Speed Sport Radar. The subjects’ perception of a change in bowling speed post-intervention was also recorded. SPSS version 15.0 was used to analyse the data. Two-tailed tests were used in all cases. Results: Trunk flexion and lateral flexion increased significantly (p < 0.05) post-spinal manipulation. There was a significant increase in bowling speed post-thoracic (p = 0.042) and post-combined manipulation (p < 0.000). A significant yet weak positive correlation (p = 0.003; r = 0.451) was seen in change in bowling speed and change in thoracic flexion and lateral flexion. There was no significant difference in the percentage subjective change by intervention group (p = 0.217). Conclusions: Spinal manipulation is a valid intervention for short-term increase in bowling speed.

碩士
國立雲林科技大學
營建工程系碩士班
93
Full-span Precast Launching Method (FPLM) is a brand new construction technic that imported by Taiwan High Speed Rail Co.. It is produced by pre-concreted completely so that the FPLM is not affected by the weather during the production. However it is depend on the substructure construction/curing completed before the deck launched that FPLM could be developed smoothly or not. You can find out it is the major temporary work, such as formwork, rebar jig, equipment etc. during the FPLM construction, in which the cost of the inner mould and outer form of box girder, the steel form of pier and pierhead of substructure, the rebar jig as well as the launching equipments are quite high. The quantity of those equipments was always depended on experience in the pass time. Therefore the more of the form quantity is used the higher of direct costs would be spent. On the other hand it would be reduced the capacity of production if the equipments are insufficient. What we concern is how to choose the best combination from thousands of different arrangements, that can both increase the production capacity and reduce the direct cost. The research is focus on the supply behavior from the manufacture and production of substructure and box girder to the precast yard and the material supplier under the satisfaction of safety stock of precast yard. Besides, the research also used software of SIMPROCESS for dynamic simulation tool to display the production process and supply behavior, Further more, the research also try to find out major affected factor of whole system to compare the difference between span numbers and progress. Finally, the research will try to use the solution of optimization for the purpose of quick searching of solution to find the best combination of resource and cost of FPLM for the decision of strategic importance of user in the future.