We continue our discussion of Gray code by presenting algorithms used to convert between the weighted numeral system of unsigned binary and the Gray code ordered sequence. We also show how to implement these algorithms in our code.

<p>Data is considered the "new oil" in the information society and digital economy. While many commercial activities and government decisions are based on data, the public raises more concerns about privacy leakage when their private data are collected and used. In this dissertation, we investigate the privacy risks in settings where the data are distributed across multiple data holders, and there is only an untrusted central server. We provide solutions for several problems under this setting with a security notion called differential privacy (DP). Our solutions can guarantee that there is only limited and controllable privacy leakage from the data holder, while the utility of the final results, such as model prediction accuracy, can be still comparable to the ones of the non-private algorithms.</p> <p><br></p> <p>First, we investigate the problem of estimating the distribution over a numerical domain while satisfying local differential privacy (LDP). Our protocol prevents privacy leakage in the data collection phase, in which an untrusted data aggregator (or a server) wants to learn the distribution of private numerical data among all users. The protocol consists of 1) a new reporting mechanism called the square wave (SW) mechanism, which randomizes the user inputs before sharing them with the aggregator; 2) an Expectation Maximization with Smoothing (EMS) algorithm, which is applied to aggregated histograms from the SW mechanism to estimate the original distributions.</p> <p><br></p> <p>Second, we study the matrix factorization problem in three federated learning settings with an untrusted server, i.e., vertical, horizontal, and local federated learning settings. We propose a generic algorithmic framework for solving the problem in all three settings. We introduce how to adapt the algorithm into differentially private versions to prevent privacy leakage in the training and publishing stages.</p> <p><br></p> <p>Finally, we propose an algorithm for solving the k-means clustering problem in vertical federated learning (VFL). A big challenge in VFL is the lack of a global view of each data point. To overcome this challenge, we propose a lightweight and differentially private set intersection cardinality estimation algorithm based on the Flajolet-Martin (FM) sketch to convey the weight information of the synopsis points. We provide theoretical utility analysis for the cardinality estimation algorithm and further refine it for better empirical performance.</p>

Traction inverter, as a critical component in electrified transportation, has been the subject of many research studies in terms of topologies, modulation, and control schemes. Recently, some of the well-known electric vehicle manufacturers have utilized higher-voltage batteries to benefit from lower current, higher power density, and faster charging times. With the ongoing trend toward higher voltage DC-link in electric vehicles, some multilevel structures have been investigated as a feasible and efficient option for replacing the two-level inverters. Higher efficiency, higher power density, better waveform quality, and inherent fault-tolerance are the foremost advantages of multilevel inverters which make them an attractive solution for this application. The first contribution of this thesis is to investigate and present a comprehensive review of the multilevel structures in traction applications. Secondly, this thesis proposes an electro-thermal model based on foster equivalent thermal networks for a designed three-level active neutral point clamped (ANPC) inverter, as well as a modified sinusoidal pulse-width modulation (SPWM) -based technique. This electro-thermal model and the modulation technique enable temperature estimation in the inverter and minimization of the hotspot temperature and hence, increase the power density. Based on the experimental results derived from the implemented setup, a 12% increase in power density is achieved with the proposed technique. The other contribution is a reduced-complexity model-predictive controller (MPC) for the three-level ANPC inverter without weighting factors in which the number of calculations has dropped from 27 to 12 in each sampling period. The improvements to the structure and control system of the inverter are supported by theoretical analysis, simulation results, and experimental tests. A three-level inverter is implemented for 800 V, 70 kW operation and tested. 750 V Silicon Carbide (SiC) switches are utilized in the inverter structure. Finally, future trends and suggestions for the following studies are stated in this thesis. / Thesis / Doctor of Philosophy (PhD)

<p> The properties and device applications of silicon thin films vacuum evaporated both onto single crystal silicon and onto silicon dioxide substrates have been investigated. </p> <p> The feasibility of obtaining device quality homoepitaxial silicon thin films by vacuum evaporation onto non heat-treated substrates having temperatures of 700°C has been demonstrated. A new technique, that of gas-doping, has been developed and has been shown to be capable of reproducibly introducing controlled concentrations of doping impurities in the range applicable to device fabrication into the deposited layers. The combined deposition-doping technique has been employed in the production of silicon layers containing impurity steps more abrupt than may be obtained by conventional fabrication techniques. </p> <p> The electrical properties of the vacuum evaporated homoepitaxial silicon layers have been shown to be comparable in most respects to those of bulk high purity single crystal silicon. The characteristics of rectifying and of varactor diodes prepared by the technique of vacuum evaporation combined with gas doping have been considered. </p> <p> Silicon films evaporated onto Si02 substrates have been shown to possess structures ranging from amorphous through randomly oriented polycrystalline to oriented polycrystalline as the substrate temperature is increased from 25°C to 850°C. The electrical characteristics of doped polycrystalline films obtained both by vacuum evaporation combined with gas doping and by the diffusion-annealing of amorphous films have been shown to be comparable with those reported for similar material deposited by chemical techniques. The experimentally observed properties of the disordered material have been qualitatively explained employing an inhomogeneous film model. The suitability of thin films of doped polycrystalline silicon on sio2 substrates for the production of high value resistors for monolithic integrated circuits has been considered. </p> / Thesis / Doctor of Philosophy (PhD)

<p> Connectionist models of language acquisition typically have difficulty with systematicity, or the ability for the network to generalize its limited experience with language to novel utterances. In this way, connectionist systems learning grammar from a set of example sentences tend to store a set of specific instances, rather than a generalized abstract knowledge of the process of grammatical combination. Further, recent models that do show limited systematicity do so at the expense of simultaneously storing explicit lexical knowledge, and also make use of both developmentally-implausible training data and biologically-implausible learning rules. Consequently, this research program develops a novel unsupervised neural network architecture, and applies this architecture to the problem of systematicity in language models.</p> <p> In the first of several studies, a connectionist architecture capable of simultaneously storing explicit and separate representations of both conceptual and grammatical information is developed, where this architecture is a hybrid of both a self-organizing map and an intra-layer Hebbian associative network. Over the course of several studies, this architecture's capacity to acquire linguistic grammar is evaluated, where the architecture is progressively refined until it is capable of acquiring a benchmark grammar consisting of several difficult clausal sentence structures - though it must acquire this grammar at the level of grammatical category, rather than the lexical level.</p> <p> The final study bridges the gap between the lexical and grammatical category levels, and develops an activation function based on a semantic feature co-occurrence metric. In concert with developmentally-plausible sensorimotor grounded conceptual representations, it is shown that a network using this metric is able to undertake a process of semantic bootstrapping, and successfully acquire separate explicit representations at the level of the concept, part-of-speech category, and grammatical sequence. This network demonstrates broadly systematic behaviour on a difficult test of systematicity, and extends its knowledge of grammar to novel sensorimotor-grounded words.</p> / Thesis / Doctor of Philosophy (PhD)

<p> In part I of this dissertation, several cationic polyacrylamides were tested under different conditions for their ability to improve the retention of fines in papermaking. A dynamic drainage jar was used to simulate the turbulence encountered in the papermaking process. Several factors, including temperature, the amount and intensity of turbulence, the additive concentration and the presence of impurities were found to affect fines retention with polymers present. A polymer made by Nalco Chemicals proved to be superior to a commonly used polymer, Percol 292 for a standard fine paper pulp. It was thought that further retention improvements might be possible by tailoring the charge density and molecular weight of polyacylamide retention aides for the specific papermaking system they are intended for. Novel approaches to retention such as those employing combinations of an anionic polymer, a cationic polymer and zirconium oxychloride were thought to show promise as well.</p> <p> In part II of this dissertation several broad polyacrylamide molecular weight standards were prepared by inverse suspension and solution processes on pilot plant equipment at the McMaster Institute for Polymer Production Technology. They were characterized by laser light scattering and viscometry at McMaster, and externally by other methods. Although the polyacrylamides prepared compare favourably to currently available commercial standards when both are analysed by SEC, further analysis must be done to be certain of the molecular weight averages.</p> <p> A relationship is presented to provide for simpler and more accurate light scattering analysis in the future. This relationship relating Mw to the second Virial coefficient may be used to eliminate some uncertainty in the often scattered plots encountered when calculating molecular weights for polyacrylamides analysed by light scattering.</p> / Thesis / Master of Engineering (MEngr)

<p> A finiteness criterion is given for finitely generated positively ordered semigroups and this is used to show that various semigroups of operators in universal algebra are finite.</p> / Thesis / Master of Science (MSc)

Catalogs of seismic source parameters (hypocenter locations, origin times, and magnitudes) are vital for studying various Earth processes, greatly enhancing our understanding of the nature of seismic events, the structure of the Earth, and the dynamics of fault systems. Modern seismic analyses utilize supervised machine learning (ML) to build enhanced catalogs based on millions of examples of analyst-picked phase-arrivals in waveforms, yet the ability to characterize the time-varying spectral content of the waveforms underlying those catalogs remains lacking. Unsupervised machine learning (UML) methods provide powerful tools for inferring patterns from musical spectrograms with little a priori information, yet has been relatively underutilized in the field of seismology. In this thesis, I leverage advanced tools from UML to analyze the temporal spectral content of large sets of spectrograms generated by different mechanisms in two distinct geologic settings: icequakes and tremors at Gornergletscher (a Swiss temperate glacier) and repeating earthquakes from a 10-km-long creeping segment of the San Andreas Fault. The core algorithm in this work, now known as Spectral Unsupervised Feature Extraction, or SpecUFEx, extracts time-varying frequency patterns from spectrograms and reduces them into low-dimensionality fingerprints via a combination of non-negative matrix factorization and hidden Markov Modeling (Holtzman et al. 2018), optimized for large data sets via stochastic variational inference. This work describes the SpecUFEx algorithm and the suite of preprocessing, clustering, and visualization tools developed to create an UML workflow, SpecUFEx+, that is widely-accessible and applicable for many seismic settings. I apply theSpecUFEx+ workflow to single- and multi-station seismic data from Gornergletscher, and demonstrate how some fingerprint-clusters track diurnal tremor related to subglacial water flow, while others correspond to the onset of the subglacial and englacial components of a glacial lake outburst flood. I also discover periods of harmonic tremor localized near the ice-bed interface that may be related to glacial stick-slip sliding. I additionally apply the SpecUFEx+ workflow to earthquakes on the San Andreas Fault to unveil far more repeating earthquake sequences than previously inferred, leading to enhanced slip-rate estimates at seismogenic depths and providing a more detailed image of seismic gaps along the fault interface. Unsupervised feature extraction is a novel tool to the field of seismology. This work demonstrates how scientific insight can be gained through the characterization of the spectral-temporal patterns of large seismic datasets within an UML-framework.

Research interest in Additive Manufacturing (AM) as an enabling technology for customizable parts is rapidly expanding. While much AM research focus is on high performance feedstocks and process optimization to obtain parts with improved mechanical properties, interest in the environmental applications of AM has recently increased. The lower cost and greater accessibility AM is leading to novel environmental research solutions in wastewater treatment and toxicity reduction by capitalizing on the increased affordability and accessibility of 3D printing (3DP) technologies for customizable, high surface area structures. The novelty and focus of this dissertation is exploration of Material Extrusion (MatEx) based Fused Filament Fabrication (FFF) of filled polymer composites as a disruptive technology enabler for deployable and retrievable structures in environmental media for adsorption, destruction and toxicity reduction of harmful chemicals. This dissertation addresses research questions that generally answer, "why AM for environmental applications?". The inherent layer-by-layer design provides larger surface area structures for interaction with contaminated media. Polylactic acid (PLA) was selected due to its green sources and biocompatibility relative to synthetic polymers and its wide processing window allowing shear thinning and "printability" despite the elevated viscosity and modulus of highly filled composites. The filler selected for contaminant adsorption was microporous zeolite, which has affinity for ammonia, radionuclides and Per- and Polyfluorinated Substances (PFAS). The filler selected for contaminant destruction was photocatalytic TiO2 nanoparticles which can degrade organic chemicals, harmful algal bloom toxins and PFAS. A preliminary research hurdle was overcome by demonstrating that immobilization of zeolite and TiO2 in a PLA binder matrix did not prevent adsorption or free radical release, respectively. The first major research objective involved investigation of high surface area printed PLA-zeolite geometries with different zeolite loadings and found that while ammonia was reduced, there were diminishing returns with increased loading in terms of mass standardized adsorptive performance due to insufficiently exposed zeolite. The research solution leveraged AM print process parameters to increase the macroporosity of the printed composite structure to create voids and channels allowing water infiltration and chemical adsorption to zeolite. Faster printing of larger roadways generated macrostructural voids that were maintained by extrusion at lower temperature for rapid solidification. The second research objective involved compounding different loadings and dispersion states of TiO2 in PLA to demonstrate immobilization of TiO2 closer to UV-light penetration water improves photocatalysis. Higher 32% w/w TiO2 loadings were heavily agglomerated and more difficult to print process due to high viscosity, rapid liquid-solid transition (G'>G") and particle network recovery during printer retractions, leading to nozzle clogging. Lower 20% w/w loading was more conducive to larger production printing due to lower viscosity, longer viscosity recovery times for retractions and thus generally a wider processing window. While altering twin screw processing parameters reduced TiO2 agglomerates in filaments, leading to increases in crystallinity (due to seeding effects and chain scission) and lower viscosity recovery, photocatalytic performance was not significantly improved. Evidence presented showed that larger particle agglomerates were more toward the inside of printed surfaces and thus less available to UV-light irradiation. This location of larger particles is supported by previous theoretical and empirical investigations showing larger particles migrate at a faster velocity away from the outer walls of confined extrudates within non-Newtonian flow fields due to normal forces, leaving more smaller particles toward outer surfaces. This research provided novel contributions to the environmental and AM research communities and pioneered a convergence of these fields into an interdisciplinary community of practice focused on better characterization and processing in environmental applications to improve structure-environmental property relationships. Future research should build on these findings to enhance performance through multi-functional materials that adsorb and destroy contaminants. The reactive surface area should be further increased through by high surface area designs and print parameter optimized porous structures providing a continuum of meso- to microporosity as confirmed by chemical flux and mass transfer studies for additional AM technologies (e.g., Direct Ink Write). / Doctor of Philosophy / Engineers and hobbies alike have great interest in Additive Manufacturing (AM), or 3D Printing, to customize parts and new designs. More recently, environmental scientists and engineers have turned to 3D printing to solve environmental problems due to the lower cost and user-friendliness of desktop machines. This research dissertation focuses on how 3D printing can allow for iterative improvements in customizable, high surface area structures to reduce chemical concentrations in water by either adsorbing or destroying the chemicals. Water is clearly a critical resource for ecosystems, recreation and drinking supplies as national security, human and ecosystem health are tied to clean water. This research addresses why 3D printing is interesting and effective for environmental solutions. Briefly the layer-by-layer design provides larger surface area structures for interaction with contaminated media. The common 3D printer feedstock Polylactic Acid (PLA) was selected since it is non-toxic and can be relatively easy to print even if modified by adding rigid filler particles for research. Micron-scale (zeolite) and nano-scale (Titanium Dioxide) particles were mixed with the polymer to make printable filaments to adsorb and destroy contaminants, respectively. This research demonstrated the proof-of-concept by removing ammonia, methylene blue dye and a harmful algal toxin from water. The materials produced are also applicable to both conventional organic pollutants and emerging contaminants of concern in the popular news such as Per- and Polyfluorinated Substances (PFAS), which were used as flame retardants and non-stick surfaces. This research ties the material properties of the experimental micro- and nano-composite filaments to how the materials extrude and solidify during 3D printing and how well the resulting printed structures work for reducing contaminant levels in water. Altering the parameters and conditions at which these materials are processed and 3D printed can significantly change their structure, density, porosity and distribution of particles and in turn increase effectiveness. The results provide new contributions to both the environmental and AM research communities and pioneers interdisciplinary collaborative ideas for these different subject matter experts to work together to better understand how handling and processing of these materials can improve their performance in environmental applications. New work should leverage the ideas and principles presented here to further improve performance, ease of production and scale-up of multifunctional material structures for multiple classes of chemicals that are of concern in surface and drinking water.

With the continuing spread of misinformation and disinformation online, it is of increasing importance to develop combating mechanisms at scale in the form of automated systems that can find checkworthy information, detect fallacious argumentation of online content, retrieve relevant evidence from authoritative sources and analyze the veracity of claims given the retrieved evidence. The robustness and applicability of these systems depend on the availability of annotated resources to train machine learning models in a supervised fashion, as well as machine learning models that capture patterns beyond domain-specific lexical clues or genre-specific stylistic insights. In this thesis, we investigate the role of models for argument structure and argument quality in improving tasks relevant to fact-checking and furthering our understanding of misinformation and disinformation. We contribute to argumentation mining, misinformation detection, and fact-checking by releasing multiple annotated datasets, developing unified models across datasets and task formulations, and analyzing the vulnerabilities of such models in adversarial settings. We start by studying the argument structure's role in two downstream tasks related to fact-checking. As it is essential to differentiate factual knowledge from opinionated text, we develop a model for detecting the type of news articles (factual or opinionated) using highly transferable argumentation-based features. We also show the potential of argumentation features to predict the checkworthiness of information in news articles and provide the first multi-layer annotated corpus for argumentation and fact-checking. We then study qualitative aspects of arguments through models for fallacy recognition. To understand the reasoning behind checkworthiness and the relation of argumentative fallacies to fake content, we develop an annotation scheme of fallacies in fact-checked content and investigate avenues for automating the detection of such fallacies considering single- and multi-dataset training. Using instruction-based prompting, we introduce a unified model for recognizing twenty-eight fallacies across five fallacy datasets. We also use this model to explain the checkworthiness of statements in two domains. Next, we show our models for end-to-end fact-checking of statements that include finding the relevant evidence document and sentence from a collection of documents and then predicting the veracity of the given statements using the retrieved evidence. We also analyze the robustness of end-to-end fact extraction and verification by generating adversarial statements and addressing areas for improvements for models under adversarial attacks. Finally, we show that evidence-based verification is essential for fine-grained claim verification by modeling the human-provided justifications with the gold veracity labels.