Automated visual inspection is an emerging multi-disciplinary field with many challenges; it combines different aspects of computer vision, pattern recognition, automation, and control systems. There does not exist a large body of work dedicated to the design of generalized visual inspection systems; that is, those that might easily be made applicable to different product types. This is an important oversight, in that many improvements in design and implementation times, as well as costs, might be realized with a system that could easily be made to function in different production environments. <br /><br /> This thesis proposes a framework for generalizing and automating the design of the defect classification stage of an automated visual inspection system. It involves using an expandable set of features which are optimized along with the classifier operating on them in order to adapt to the application at hand. The particular implementation explored involves optimizing the feature set in disjoint sets logically grouped by feature type to keep search spaces reasonable. Operator input is kept at a minimum throughout this customization process, since it is limited only to those cases in which the existing feature library cannot adequately delineate the classes at hand, at which time new features (or pools) may have to be introduced by an engineer with experience in the domain. <br /><br /> Two novel methods are put forward which fit well within this framework: cluster-space and hybrid-space classifiers. They are compared in a series of tests against both standard benchmark classifiers, as well as mean and majority vote multi-classifiers, on feature sets comprised of just the logical feature subsets, as well as the entire feature sets formed by their union. The proposed classifiers as well as the benchmarks are optimized with both a progressive combinatorial approach and with an genetic algorithm. Experimentation was performed on true colour industrial lumber defect images, as well as binary hand-written digits. <br /><br /> Based on the experiments conducted in this work, it was found that the sequentially optimized multi hybrid-space methods are capable of matching the performances of the benchmark classifiers on the lumber data, with the exception of the mean-rule multi-classifiers, which dominated most experiments by approximately 3% in classification accuracy. The genetic algorithm optimized hybrid-space multi-classifier achieved best performance however; an accuracy of 79. 2%. <br /><br /> The numeral dataset results were less promising; the proposed methods could not equal benchmark performance. This is probably because the numeral feature-sets were much more conducive to good class separation, with standard benchmark accuracies approaching 95% not uncommon. This indicates that the cluster-space transform inherent to the proposed methods appear to be most useful in highly dependant or confusing feature-spaces, a hypothesis supported by the outstanding performance of the single hybrid-space classifier in the difficult texture feature subspace: 42. 6% accuracy, a 6% increase over the best benchmark performance. <br /><br /> The generalized framework proposed appears promising, because classifier performance over feature sets formed by the union of independently optimized feature subsets regularly met and exceeded those classifiers operating on feature sets formed by the optimization of the feature set in its entirety. This finding corroborates earlier work with similar results [3, 9], and is an aspect of pattern recognition that should be examined further.

The quantitative analysis of decomposed electromyographic (EMG) signals reveals information for diagnosing and characterizing neuromuscular disorders. Neuromuscular jitter is an important measure that reflects the stability of the operation of a neuromuscular junction. It is conventionally measured using single fiber electromyographic (SFEMG) techniques. SFEMG techniques require substantial physician dexterity and subject cooperation. Furthermore, SFEMG needles are expensive, and their re-use increases the risk of possible transmission of infectious agents. Using disposable concentric needle (CN) electrodes and automating the measurment of neuromuscular jitter would greatly facilitate the study of neuromuscular disorders. An improved automated jitter measurment system based on the decomposition of CN detected EMG signals is developed and evaluated in this thesis. Neuromuscular jitter is defined as the variability of time intervals between two muscle fiber potentials (MFPs). Given the candidate motor unit potentials (MUPs) of a decomposed EMG signal, which is represented by a motor unit potential train (MUPT), the automated jitter measurement system designed in this thesis can be summarized as a three-step procedure: 1) identify isolated motor unit potentials in a MUPT, 2) detect the significant MFPs of each isolated MUP, 3) track significant MFPs generated by the same muscle fiber across all isolated MUPs, select typical MFP pairs, and calculate jitter. In Step one, a minimal spanning tree-based 2-phase clustering algorithm was developed for identifying isolated MUPs in a train. For the second step, a pattern recognition system was designed to classify detected MFP peaks. At last, the neuromuscular jitter is calculated based on the tracked and selected MFP pairs in the third step. These three steps were simulated and evaluated using synthetic EMG signals independently, and the whole system is preliminary implemented and evaluated using a small simulated data base. Compared to previous work in this area, the algorithms in this thesis showed better performance and great robustness across a variety of EMG signals, so that they can be applied widely to similar scenarios. The whole system developed in this thesis can be implemented in a large EMG signal decomposition system and validated using real data.

In this thesis, I have studied CMOS dendrites, implemented them on a reconfigurable analog platform and modeled them using MATLAB Simulink. The dendrite model was further used to build a computational model. I implemented a Hidden Markov Model (HMM) classifier to build a simple YES/NO wordspotter. I also discussed the inter-relation between neural systems, CMOS transistors and HMM networks. The physical principles behind the operation of silicon devices and biological structures are similar. Hence silicon devices can be used to emulate biological structures like dendrites. Dendrites are a branched, conductive medium which connect a neurons synapses to its soma. Dendrites were previously believed to be like wires in neural networks. However, recent research suggests that they have computational power. We can emulate dendrites using transistors in the Field Programmable Analog Array (FPAA). Our lab has built the Reconfigurable Analog Signal Processor (RASP) family of FPAAs which was used for the experiments. I analytically compared the mathematical model of dendrites to our model in silicon. The mathematical model based on the device physics of the silicon devices was then used to simulate dendrites in Simulink. An automated tool, sim2spice was then used to convert the Simulink model into a SPICE netlist, such that it can be implemented on a FPAA. This is an easier tool to use for DSP and Neuromorphic engineers who's primary areas of expertise isn't circuit design.

This thesis concerns Social Network Analysis as a mechanism for exploring Community Dynamics. To be able to use the Social Network methodologies, relationships existing between the modeling entities are required. In this thesis, we use two different kinds of relationships: e-mails exchanged and co-authorship of papers. The e-mails exchanged, as an indicator of information exchange in an organization, is used to facilitate the emergence of structure within the organization. In this thesis we demonstrate the effectiveness of using e-mail communication patterns for crisis detection in a hierarchically set organization. We compare the performance of a Social Network based Classifier with some of the traditional classifiers from the data mining framework for inferring this hierarchy. A generic framework for studying dynamic group transformations is presented and the co-authorship of papers, as an indicator of collaboration in an academic institution, is used to study the community behavioral patterns evolving over time. Enron e-mail corpus and the IISc Co-authorship Dataset are utilized for illustrative purposes.

The United States Navy is committed to developing technology for an All-Electric Ship (AES) that promises to improve the affordability and capability of its next-generation warships. With the addition of power-intensive 21st century electrical systems, future thermal loads are projected to exceed current heat removal capacity. Furthermore, rising fuel costs necessitate a careful approach to total-ship energy management. Accordingly, the aim of this research is to develop computer tools for early-stage design of shipboard energy distribution systems. A system-level model is developed that enables ship designers to assess the effects of thermal and electrical system configurations on fuel efficiency and survivability. System-level optimization and design exploration, based on these energy system models, is challenging because the models are sometimes computationally expensive and characterized by discrete design variables and discontinuous responses. To address this challenge, a classifier-guided sampling (CGS) method is developed that uses a Bayesian classifier to pursue solutions with desirable performance characteristics. The CGS method is tested on a set of example problems and applied to the AES energy system model. Results show that the CGS method significantly improves the rate of convergence towards known global optima, on average, when compared to genetic algorithms. / text

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal

Software engineers use design patterns to refactor software models for quality. This displaces domain patterns and makes software hard to maintain. Detecting design patterns directly in requirements can circumvent this problem. To facilitate the analogical transfer of patterns from problem domain to solution model however we must describe patterns in ontological rather than in technical terms. In a first study novice designers used both pattern cases and a pattern ontology to detect design ideas and patterns in requirements. Errors in detection accuracy led to the revision of the pattern ontology and a second study into its pattern-discriminating power. Study results demonstrate that pattern ontology is superior to pattern cases in assisting novice software engineers in identifying patterns in the problem domain.

自Greenberg (1972) 以來，陸續有學者探討世界語言中的複數標記和分類詞之間的關係，並指出兩者的相似之處(Greenberg, 1972; Sanches and Slobin, 1973; Borer, 2005; Her, 2012)。本文稱這些研究為CL-PM Convergence View。這些研究指出分類詞 (numeral classifier; CL) 與複數標記 (plural marker; PM)呈現互補分布的關係。學者們 (Borer, 2005; Her, 2012) 更進一步認為，分類詞與複數標記在同一名詞組裡呈現互補分布，表示兩者在句法結構上佔據相同的位置，應視為相同的成分。然而本文發現，台灣華語「Num+CL+N們」結構有一定的能產性。若將「們」視為複數標記，則台灣華語對於上述學者的理論就形成了反例。有鑑於此，本文目的在於以句法接受度實驗及語料庫兩項方法，重新檢視台灣華語「們」的各種用法，以期解決文獻上對於「們」的諸多爭議。接著，在建立語言事實後，本文探討「們」對於CL-PM Convergence View的意義，並對於台灣華語「們」與分類詞在歷時和共時上的互動作出新的解釋。 具體來說，本文釐清了下列六項事實，為文獻上的爭議提供新的證據: (一)「們」可用於非指人的名詞；(二)「N們」為定指；(三)「Proper N們」僅表示 ‘多位Proper N’；(四)「¬1群N們」合法；(五)「Num+CL+N們」合法；(六) 在有接受英語教育的前提下，母語者的英語程度越低，對於「Num+CL+N們」的接受度越高，也越容易受到英語句法結構的促發(priming)，表示「Num+CL+N們」的產生與和英語的接觸有密切關係。考慮這些事實及其他台灣華語沒有爭議的特性，本文認為台灣華語「們」應視為一個集合標記(collective plural marker)，而非文獻所說的伴同標記(associative plural, Iljic, 2001,2005; 陳俊光，2009)或是普通複數標記(additive plural, Li, 1999; Hunag et al, 2009)。最後，本文提出兩個論點: 第一，我們根據Her et al (to appear)的洞見，區分「語意複數」(semantic plural)及「語法複數」(grammatical plural)；第二，我們提出新的事實，論證「們」在句法上，台灣華語「們」是一個附綴(clitic)而非詞綴(suffix)。這兩項論點證明台灣華語「們」並不違反CL-PM Convergence View的預測，亦可以解釋「們」與分類詞在歷時(李豔惠、石毓智，2000)和共時上的互動。 / Since Greenberg (1972), there have been many studies addressing the issue of the relationship between numeral classifiers (CL) and plural markers (PM) (Greenberg, 1972; Sanches and Slobin, 1973; Borer, 2005; Her, 2012). These scholars (henceforth CL-PM Convergence View) discovered that CL and PM tend not to co-occur in the same language, and even if they do co-occur, they are complementarily distributed within NP. Some linguists (Borer, 2005; Her, 2012) take this generalization further to propose that CL and PM in fact belong to the same category. However, when we look at Taiwan Mandarin (TM) data, a potential counterevidence can be found: [Num+CL+N-men], in which CL and –men, generally taken to be a plural suffix, co-occur within the same NP. In light of this, this study aims to take a realist look at TM –men, collecting relevant data from grammaticality judgment task and corpora so as to capture the behavior of –men. We then test CL-PM Convergence View against empirical data obtained in the study, showing that [Num+CL+N-men] does not constitute a counterexample to CL-PM Convergence View. The apparent interaction between CL and –men in TM can also be accounted for under our analysis of –men. Specifically, this study establishes the following facts for TM –men: (1) the use of –men is not restricted to human Ns; (2) N-men must be definite; (3) Proper N denotes ‘more than one Proper N’; (4) [¬1 qun N-men] is grammatical; (5) [Num+CL+N-men] is grammatical; (6) native speakers’ acceptability of [Num+CL+N-men] is in negative correlation with their English proficiency, and priming effects of English structure [Num+N-s] are observed on speakers with low English proficiency. Taking these findings into account, this study proposes that TM –men should be best analyzed as a collective plural marker, contra Iljic, (2001,2005) and 陳俊光’s (2009) “associative” analysis on the one hand, and Li (1999) and Hunag et al’s (2009) “additive” analysis on the other. Accordingly, we argue that –men as a collective does not constitute a counterexample to CL-PM Convergence View, citing two further pieces of evidence: Her et al’s (to appear) insight that “semantic plural” and “grammatical plural” should be distinguished and the proposal made there to revise CL-PM Convergence View, and the “clitic” analysis of TM –men proposed in this study. Finally, we show that the distinction between “semantic plural” and “grammatical plural” also nicely explains the synchronic and diachronic interaction between CL and –men in TM.

本論文的主要目的在對於分類詞、性別詞以及複數標記在語言當中的地域分佈提出解釋．在前人的研究當中，這三項元素被認為是名詞句中平衡資訊的重要工具(Greenberg, 1990; Aikhenvald, 2000)：分類詞語言主要位於東南亞和南美洲部分地區，而具有性別詞或複數標記的語言大多出現在歐洲、非洲和美洲部分地區．我們提出的論證如下：即便這三樣元素外表上具有歧異，它們會呈現當今所見的地域分佈原因在於它們共有的兩項標記功能：可數性質及語意分類．分類詞同時滿足兩者而性別詞及複數標記分別滿足其一；依照此邏輯，我們預測有分類詞的語言不會同時具有性別詞及複數標記而反之亦然．本文中我們透過句法形式和語意功能的比較提出論證並透過類型學、地理及歷史的角度分析來自世界上最大的二十個語系（印歐，漢藏，亞非，尼日爾-剛果，南島，達羅毗荼，阿爾泰，南亞，壯侗，尼羅－撒哈拉，烏拉，高加索，等語系）的155個語言．架構上，第一章簡單對研究題目進行介紹，第二章呈現前人研究的匯整，第三章包含我們的理論論證以及我們對於分類詞、性別詞及複數標記分與合的解釋．隨後的第四章中，我們提出類型學和地理資訊系統(GIS)的證據；最後在第五張和第六章我們分別點出本研究的限制以及結論． / This thesis aims at providing an explanation for the typological and areal distribution between numeral classifiers, genders (noun classes) and grammatical plural markers. Within previous studies, these three components are considered as different devices to balance information in noun phrases (Greenberg, 1990; Aikhenvald, 2000). Numeral classifier languages are mainly present in South-East Asia and parts of South-America, while languages with genders and grammatical plural markers are generally attested in Europe, Africa and parts of the Americas. We propose that despite their apparent divergence, the three elements display this particular geographical distribution due to their convergent features of count/mass distinction and semantic classification: Numeral classifiers carry both functions, while genders and plural markers separately fulfill one of them. Following this logic, we expect that a language with numeral classifier do not have simultaneously the systems of genders plus plural markers and vice-versa. Theoretical evidence via formal syntactic form and semantic function comparison is proposed and further supported by typological, geographical and historical analysis of 155 languages that are mainly part of the 20 biggest language groups in the world, e.g. Indo-European, Sino-Tibetan, Afro-Asiatic, Niger-Congo, Austronesian, Dravidian, Japonic, Altaic, Austro-Asiatic, Tai-Kadai, Creole, Nilo-Saharan, Uralic, Quechuan, Hmong-Mien, Mayan, North Caucasian, Language isolates among others. Chapter 1 presents a brief introduction of the subject while chapter 2 displays the literature review. Chapter 3 includes our theoretical discussion proposing explaining the convergence and divergence among numeral classifiers, genders and plural markers, followed by typological and geographical evidence via GIS (Geographic Information System) in Chapter 4. Finally Chapter 5 and 6 contain the limitations of our study and its conclusion.

La maintenance du logiciel est une phase très importante du cycle de vie de celui-ci. Après les phases de développement et de déploiement, c’est celle qui dure le plus longtemps et qui accapare la majorité des coûts de l'industrie. Ces coûts sont dus en grande partie à la difficulté d’effectuer des changements dans le logiciel ainsi que de contenir les effets de ces changements. Dans cette perspective, de nombreux travaux ont ciblé l’analyse/prédiction de l’impact des changements sur les logiciels. Les approches existantes nécessitent de nombreuses informations en entrée qui sont difficiles à obtenir. Dans ce mémoire, nous utilisons une approche probabiliste. Des classificateurs bayésiens sont entraînés avec des données historiques sur les changements. Ils considèrent les relations entre les éléments (entrées) et les dépendances entre changements historiques (sorties). Plus spécifiquement, un changement complexe est divisé en des changements élémentaires. Pour chaque type de changement élémentaire, nous créons un classificateur bayésien. Pour prédire l’impact d’un changement complexe décomposé en changements élémentaires, les décisions individuelles des classificateurs sont combinées selon diverses stratégies. Notre hypothèse de travail est que notre approche peut être utilisée selon deux scénarios. Dans le premier scénario, les données d’apprentissage sont extraites des anciennes versions du logiciel sur lequel nous voulons analyser l’impact de changements. Dans le second scénario, les données d’apprentissage proviennent d’autres logiciels. Ce second scénario est intéressant, car il permet d’appliquer notre approche à des logiciels qui ne disposent pas d’historiques de changements. Nous avons réussi à prédire correctement les impacts des changements élémentaires. Les résultats ont montré que l’utilisation des classificateurs conceptuels donne les meilleurs résultats. Pour ce qui est de la prédiction des changements complexes, les méthodes de combinaison "Voting" et OR sont préférables pour prédire l’impact quand le nombre de changements à analyser est grand. En revanche, quand ce nombre est limité, l’utilisation de la méthode Noisy-Or ou de sa version modifiée est recommandée. / Software maintenance is one of the most important phases in the software life cycle. After the development and deployment phases, maintenance is a continuous phase that lasts until removing the software from operation. It is then the most costly phase. Indeed, those costs are due to the difficulty of implementing different changes in the system and to manage their impacts. In this context, much research work has targeted the problem of change impact analysis/prediction. The existent approaches require many inputs that are difficult to extract. In this Master thesis, we propose a probabilistic approach that uses Bayesian classifiers to predict the change impact. These classifiers are trained with historical data about changes. The consider the relations between the elements of a system (input), and the dependencies between the occurred changes (output). More precisely, a complex change in a system is divided into a set of elementary changes. For each type of elementary change, we create a classifier. To predict the impact of complex change, the individual decisions of each classifier are combined using different strategies. We evaluate our approach in two scenarios. In the first, we extract the learning data from the oldest versions of the same system. In the second scenario, the learn data comes from other systems to create the classifiers. This second scenario is interesting because it allows us to use our approach on systems without change histories. Our approach showed that it can predict the impact of elementary changes. The best results are obtained using the classifiers based on conceptual relations. For the prediction of complex changes by the combination of elementary decisions, the results are encouraging considering the few used inputs. More specifically, the voting method and the OR method predict better complex changes when the number of case to analyze is large. Otherwise, using the method Noisy-Or or its modified version is recommended when the number of cases is small.