Machine learning via transitions

Machine Learning is a modern and actively developing field of computer science, devoted to extracting and estimating dependencies from empirical data. It combines such fields as statistics, optimization theory and artificial intelligence. In practical tasks, the general aim of Machine Learning is to construct algorithms able to generalize and predict in previously unseen situations based on some set of examples. Given some finite information, Machine Learning provides ways to exract knowledge, describe, explain and predict from data. Kernel Methods are one of the most successful branches of Machine Learning. They allow applying linear algorithms with well-founded properties such as generalization ability, to non-linear real-life problems. Support Vector Machine is a well-known example of a kernel method, which has found a wide range of applications in data analysis nowadays. In many practical applications, some additional prior knowledge is often available. This can be the knowledge about the data domain, invariant transformations, inner geometrical structures in data, some properties of the underlying process, etc. If used smartly, this information can provide significant improvement to any data processing algorithm. Thus, it is important to develop methods for incorporating prior knowledge into data-dependent models. The main objective of this thesis is to investigate approaches towards learning with kernel methods using prior knowledge. Invariant learning with kernel methods is considered in more details. In the first part of the thesis, kernels are developed which incorporate prior knowledge on invariant transformations. They apply when the desired transformation produce an object around every example, assuming that all points in the given object share the same class. Different types of objects, including hard geometrical objects and distributions are considered. These kernels were then applied for images classification with Support Vector Machines. Next, algorithms which specifically include prior knowledge are considered. An algorithm which linearly classifies distributions by their domain was developed. It is constructed such that it allows to apply kernels to solve non-linear tasks. Thus, it combines the discriminative power of support vector machines and the well-developed framework of generative models. It can be applied to a number of real-life tasks which include data represented as distributions. In the last part of the thesis, the use of unlabelled data as a source of prior knowledge is considered. The technique of modelling the unlabelled data with a graph is taken as a baseline from semi-supervised manifold learning. For classification problems, we use this apporach for building graph models of invariant manifolds. For regression problems, we use unlabelled data to take into account the inner geometry of the input space. To conclude, in this thesis we developed a number of approaches for incorporating some prior knowledge into kernel methods. We proposed invariant kernels for existing algorithms, developed new algorithms and adapted a technique taken from semi-supervised learning for invariant learning. In all these cases, links with related state-of-the-art approaches were investigated. Several illustrative experiments were carried out on real data on optical character recognition, face image classification, brain-computer interfaces, and a number of benchmark and synthetic datasets.

This dissertation addresses two significant issues in the effective application of machine learning algorithms and models for the physical and engineering sciences. The first is the broad challenge of automated modeling of data across different processes in a physical system. The second is the dilemma of obtaining insightful interpretations on the relationships between the inputs and outcome of a system as inferred from complex, black box machine learning models.Automated Geometric Shape Deviation Modeling for Additive Manufacturing SystemsAdditive manufacturing systems possess an intrinsic capability for one-of-a-kind manufacturing of a vast variety of shapes across a wide spectrum of processes. One major issue in AM systems is geometric accuracy control for the inevitable shape deviations that arise in AM processes. Current effective approaches for shape deviation control in AM involve the specification of statistical or machine learning deviation models for additively manufactured products. However, this task is challenging due to the constraints on the number of test shapes that can be manufactured in practice, and limitations on user efforts that can be devoted for learning deviation models across different shape classes and processes in an AM system. We develop an automated, Bayesian neural network methodology for comprehensive shape deviation modeling in an AM system. A fundamental innovation in this machine learning method is our new and connectable neural network structures that facilitate the transfer of prior knowledge and models on deviations across different shape classes and AM processes. Several case studies on in-plane and out-of-plane deviations, regular and free-form shapes, and different settings of lurking variables serve to validate the power and broad scope of our methodology, and its potential to advance high-quality manufacturing in an AM system.Interpretable Machine LearningMachine learning algorithms and models constitute the dominant set of predictive methods for a wide range of complex, real-world processes. However, interpreting what such methods effectively infer from data is difficult in general. This is because their typical black box natures possess a limited ability to directly yield insights on the underlying relationships between inputs and the outcome for a process. We develop methodologies based on new predictive comparison estimands that effectively enable one to ``mine’’ machine learning models, in the sense of (a) interpreting their inferred associations between inputs and/or functional forms of inputs with the outcome, (b) identifying the inputs that they effectively consider relevant, and (c) interpreting the inferred conditional and two-way associations of the inputs with the outcome. We establish Fisher consistent estimators, and their corresponding standard errors, for our new estimands under a condition on the inputs' distributions. The significance of our predictive comparison methodology is demonstrated with a wide range of simulation and case studies that involve Bayesian additive regression trees, neural networks, and support vector machines. Our extended study of interpretable machine learning for AM systems demonstrates how our method can contribute to smarter advanced manufacturing systems, especially as current machine learning methods for AM are lacking in their ability to yield meaningful engineering knowledge on AM processes.

Continuing Humanity’s Dialogue with Nature and Itself

ABSTRACTMachine learning has been widely applied to clearly defined problems of astronomy and astrophysics. However, deep learning and its conceptual differences to classical machine learning have been largely overlooked in these fields. The broad hypothesis behind our work is that letting the abundant real astrophysical data speak for itself, with minimal supervision and no labels, can reveal interesting patterns that may facilitate discovery of novel physical relationships. Here, as the first step, we seek to interpret the representations a deep convolutional neural network chooses to learn, and find correlations in them with current physical understanding. We train an encoder–decoder architecture on the self-supervised auxiliary task of reconstruction to allow it to learn general representations without bias towards any specific task. By exerting weak disentanglement at the information bottleneck of the network, we implicitly enforce interpretability in the learned features. We develop two independent statistical and information-theoretical methods for finding the number of learned informative features, as well as measuring their true correlation with astrophysical validation labels. As a case study, we apply this method to a data set of ∼270 000 stellar spectra, each of which comprising ∼300 000 dimensions. We find that the network clearly assigns specific nodes to estimate (notions of) parameters such as radial velocity and effective temperature without being asked to do so, all in a completely physics-agnostic process. This supports the first part of our hypothesis. Moreover, we find with high confidence that there are ∼4 more independently informative dimensions that do not show a direct correlation with our validation parameters, presenting potential room for future studies.

Gesture recognition is a topic in computer science and language technology that aims to interpret human gestures with computer programs and many different algorithms. It can be seen as the way computers can understand human body language. Today, the main interaction tools between computers and humans are still the keyboard and mouse. Gesture recognition can be used as a tool for communication with the machine and interaction without any mechanical device such as a keyboard or mouse. In this paper, we present the results of a comparison of eight different machine learning (ML) classifiers in the task of human hand gesture recognition and classification to explore how to efficiently implement one or more tested ML algorithms on an 8-bit AVR microcontroller for on-line human gesture recognition with the intention to gesturally control the mobile robot. The 8-bit AVR microcontrollers are still widely used in the industry, but due to their lack of computational power and limited memory, it is a challenging task to efficiently implement ML algorithms on them for on-line classification. Gestures were recorded by using inertial sensors, gyroscopes, and accelerometers placed at the wrist and index finger. One thousand and eight hundred (1800) hand gestures were recorded and labelled. Six important features were defined for the identification of nine different hand gestures using eight different machine learning classifiers: Decision Tree (DT), Random Forests (RF), Logistic Regression (LR), Linear Discriminant Analysis (LDA), Support Vector Machine (SVM) with linear kernel, Naïve Bayes classifier (NB), K-Nearest Neighbours (KNN), and Stochastic Gradient Descent (SGD). All tested algorithms were ranged according to Precision, Recall, and F1-score (abb.: P-R-F1). The best algorithms were SVM (P-R-F1: 0.9865, 0.9861, and 0.0863), and RF (P-R-F1: 0.9863, 0.9861, and 0.0862), but their main disadvantage is their unusability for on-line implementations in 8-bit AVR microcontrollers, as proven in the paper. The next best algorithms have had only slightly poorer performance than SVM and RF: KNN (P-R-F1: 0.9835, 0.9833, and 0.9834) and LR (P-R-F1: 0.9810, 0.9810, and 0.9810). Regarding the implementation on 8-bit microcontrollers, KNN has proven to be inadequate, like SVM and RF. However, the analysis for LR has proved that this classifier could be efficiently implemented on targeted microcontrollers. Having in mind its high F1-score (comparable to SVM, RF, and KNN), this leads to the conclusion that the LR is the most suitable classifier among tested for on-line applications in resource-constrained environments, such as embedded devices based on 8-bit AVR microcontrollers, due to its lower computational complexity in comparison with other tested algorithms.

Background and aimsThe birth weight of a newborn is a crucial factor that affects their overall health and future well-being. Low birth weight (LBW) is a widespread global issue, which the World Health Organization defines as weighing less than 2,500 g. LBW can have severe negative consequences on an individual’s health, including neonatal mortality and various health concerns throughout their life. To address this problem, this study has been conducted using BDHS 2017–2018 data to uncover important aspects of LBW using a variety of machine learning (ML) approaches and to determine the best feature selection technique and best predictive ML model.MethodsTo pick out the key features, the Boruta algorithm and wrapper method were used. Logistic Regression (LR) used as traditional method and several machine learning classifiers were then used, including, DT (Decision Tree), SVM (Support Vector Machine), NB (Naïve Bayes), RF (Random Forest), XGBoost (eXtreme Gradient Boosting), and AdaBoost (Adaptive Boosting), to determine the best model for predicting LBW. The model’s performance was evaluated based on the specificity, sensitivity, accuracy, F1 score and AUC value.ResultsResult shows, Boruta algorithm identifies eleven significant features including respondent’s age, highest education level, educational attainment, wealth index, age at first birth, weight, height, BMI, age at first sexual intercourse, birth order number, and whether the child is a twin. Incorporating Boruta algorithm’s significant features, the performance of traditional LR and ML methods including DT, SVM, NB, RF, XGBoost, and AB were evaluated where LR, had a specificity, sensitivity, accuracy and F1 score of 0.85, 0.5, 85.15% and 0.915. While the ML methods DT, SVM, NB, RF, XGBoost, and AB model’s respective accuracy values were 85.35%, 85.15%, 84.54%, 81.18%, and 84.41%. Based on the specificity, sensitivity, accuracy, F1 score and AUC, RF (specificity = 0.99, sensitivity = 0.58, accuracy = 85.86%, F1 score = 0.9243, AUC = 0.549) outperformed the other methods. Both the classical (LR) and machine learning (ML) models’ performance has improved dramatically when important characteristics are extracted using the wrapper method. The LR method identified five significant features with a specificity, sensitivity, accuracy and F1 score of 0.87, 0.33, 87.12% and 0.9309. The region, whether the infant is a twin, and cesarean delivery were the three key features discovered by the DT and RF models, which were implemented using the wrapper technique. All three models had the identical F1 score of 0.9318. However, “child is twin” was recognized as a significant feature by the SVM, NB, and AB models, with an F1 score of 0.9315. Ultimately, with an F1 score of 0.9315, the XGBoost model recognized “child is twin” and “age at first sex” as relevant features. Random Forest again beat the other approaches in this instance.ConclusionsThe study reveals Wrapper method as the optimal feature selection technique. The ML method outperforms traditional methods, with Random Forest (RF) being the most effective predictive model for Low-Birth-Weight prediction. The study suggests that policymakers in Bangladesh can mitigate low birth weight newborns by considering identified risk factors.

Missing data problem is ubiquitous in many real life situations. Information Technology researchers have explored and tried to address this problem in different settings. In this thesis, we undertake research to address missing data problem associated with order book information in stock markets. This is an in-depth and large-scale study with systematic and comprehensive framework to address missing data problem in the finance literature. Orders placed by traders and the corresponding order imbalance (OIB) is informative to predict future stock returns, however, stock exchange rules do not reveal price sensitive complete order book data for traders. Hence, return prediction using the revealed, incomplete trade book data (that contains only matched buy and sell orders and deletes the unmatched orders), does not let traders to completely exploit possible short term trading opportunities. Hence, this can be considered as a classical missing data problem for predicting future returns, by using the information content of order book. This thesis addresses the missing data problem by developing an integrated theoretical framework applied in stock market trading environment. We use relational Markov networks theory and build an empirically testable Algorithm for Imputed Complete Order Book (AICOB). The thesis contributes by developing a new theoretical advancement of information technology research relating to missing data problem and applying it to financial markets. First, the thesis presents the missing data problem as a Missing at Random (MAR) data and builds a systematic framework to estimate single as well as joint log likelihood functions. The thesis demonstrates that estimating by using incomplete records, improves the accuracy of the parameter estimates. Second, the thesis proposes a Relational Markov Network Model for estimation of the joint distribution function of orders, order characteristics and their interactions. Later, the Expectation Maximization Algorithm is proposed to address the missing data problem during the joint estimation procedure. All pooled regression results follow Fama and MacBeth (1973) and Generalized Methods of Moments (GMM) methods. These methods control the cross sectional and time series correlations between the observations and across the pooled stocks. The proposed novel methodology overcomes the estimation problem in the context of missing order book data. Third, the thesis develops an objective evaluation strategy for AICOB, based on efficiency, accuracy and adaptability dimensions. The thesis uses Australian stock market data, which provides not only trade book data but also historical order book data for cross validating the results. This unique setting allows validating the accuracy of AICOB methodology by comparing with complete order book data. The main contribution of the thesis is to show that AICOB based predictions match with the complete order book data. Whereas, trade book based predictions are quite inconsistent to the complete order book data. The AICOB based results are also consistent with the theoretical predictions proposed in finance literature that OIB predicts better as the firm size increases. The results show that large firms, with higher trading activity and more competition for order flows, report more significant OIB prediction of future returns. Trade Book based OIB estimates, which suffer from missing data problem, fail to predict future returns for stock portfolios. Hence, addressing the missing data problem is important before implementing OIB based trading strategies. Overall, the thesis finds that machine learning applications, similar to AICOB, can be helpful in implementing the trading strategies in financial markets. The imputation of missing data, through systematic procedures, based on theoretical distributional properties of the financial variables, can be informative for more accurate prediction of future returns. The thesis contributes towards establishing a common ground for cross-disciplinary research in Finance and Information Technology (IT) by applying the advances in IT research to solve research and corresponding implementation problems in finance. Further, the thesis contributes towards advancing the order imbalance literature by showing that missing data can play a critical role in the predictive ability of order imbalance.

To understand students’ learning behaviors, this study uses machine learning technologies to analyze the data of interactive learning environments, and then predicts students’ learning outcomes. This study adopted a variety of machine learning classification methods, quizzes, and programming system logs, found that students’ learning characteristics were correlated with their learning performance when they encountered similar programming practice. In this study, we used random forest (RF), support vector machine (SVM), logistic regression (LR), and neural network (NN) algorithms to predict whether students would submit on time for the course. Among them, the NN algorithm showed the best prediction results. Education-related data can be predicted by machine learning techniques, and different machine learning models with different hyperparameters can be used to obtain better results.

Bidirectional Human-Robot Learning: Imitation and Skill Improvement

Robots had a great impact on the manufacturing industry ever since the early seventies when companies such as KUKA and ABB started deploying their first industrial robots. These robots merely performed very specific tasks in specific ways within well-defined environments. Still, they proved to be very useful as they could exceed human performance at these tasks. However, in order to enable robots to enter our daily life, they need to become more versatile and need to operate in much less structured environments. This thesis is partly devoted to stretching these limitations by means of learning, namely imitation learning (IL) and inverse reinforcement learning (IRL). Reinforcement learning (RL) is a powerful approach to enable robots to solve a task in an unknown environment. The practitioner describes a desired behavior by specifying a reward function and the robot autonomously interacts with the environment in order to find a control policy that generates high accumulated reward. However, RL is not suitable for teaching new tasks by non-experts because specifying appropriate reward functions can be difficult. Demonstrating the desired behavior is often easier for non-experts. Imitation learning can be used in order to enable the robot to reproduce the demonstrations. However, without explicitly inferring and modeling the intentions of the demonstrations, it can become difficult to solve the task for unseen situations. Inverse reinforcement learning (IRL) therefore aims to infer a reward function from the demonstrations, such that optimizing this reward function yields the desired behavior even for different situations. This thesis introduces a unifying approach to solve the inverse reinforcement learning problem in the same way as the reinforcement learning problem. This is achieved by framing both problems as information projection problems, i.e., we strive to minimize the relative entropy between a probabilistic model of the robot behavior and a given desired distribution. Furthermore, a trust region on the robot behavior is used to stabilize the optimization. For inverse reinforcement learning, the desired distribution is implicitly given by the expert demonstrations. The resulting optimization can be efficiently solved using state-of-the-art reinforcement learning methods. For reinforcement learning, the log-likelihood of the desired distribution is given by the reward function. The resulting optimization problem corresponds to a standard reinforcement learning formulation, except for an additional objective of maximizing the entropy of the robot behavior. This entropy objective adds little overhead to the optimization, but can lead to better exploration and more diversified policies. Trust-region I-projections are not only useful for training robots, but can also be applied to other machine learning problems. I-projections are typically used for variational inference, in order to approximate an intractable distribution by a simpler model. However, the resulting optimization problems are usually optimized based on stochastic gradient descent which often suffers from high variance in the gradient estimates. As trust-region I-projections where shown to be effective for reinforcement learning and inverse reinforcement learning, this thesis also explores their use for variational inference. More specifically, trust-region I-projections are investigated for the problem of approximating an intractable distribution by a Gaussian mixture model (GMM) with an adaptive number of components. GMMs are highly desirable for variational inference because they can yield arbitrary accurate approximations while inference from GMMs is still relatively cheap. In order to make learning the GMM feasible, we derive a lower bound that enables us to decompose the objective function. The optimization can then be performed by iteratively updating individual components using a technique from reinforcement learning. The resulting method is capable of learning approximations of significantly higher quality than existing variational inference methods. Due to the similarity of the underlying optimization problems, the insights gained from our variational inference method are also useful for IL and IRL. Namely, a similar lower bound can be applied also for the I-projection formulation of imitation learning. However, whereas for variational inference the lower bound serves to decompose the objective function, for imitation learning it allows us to provide a reward signal to the robot that does not depend on its behavior. Compared to reward functions that are relative to the current behavior of the robot---which are typical for popular adversarial methods---behavior-independent reward functions have the advantages that we can show convergence even for greedy optimization. Furthermore, behavior-independent reward functions solve the inverse reinforcement learning problem, thereby closing the gap between imitation learning and IRL. However, algorithms derived from our non-adversarial formulation are actually very similar to existing AIL methods, and we can even show that adversarial inverse reinforcement learning (AIRL) is indeed an instance of our formulation. AIRL was derived from an adversarial formulation, and we point out several problems of that derivation. In contrast, we show that AIRL can be straightforwardly derived from out non-adversarial formulation. Furthermore, we demonstrate that the non-adversarial formulation can be also used to derive novel algorithms by presenting a non-adversarial method for offline imitation learning.

The preparation (aqueous chemistry and the thermal activation of polyoxometalates) and the structures of promoted molybdenum oxide catalysts are investigated under the condition of selective propene oxidation by in situ Raman spectroscopy. The influences of V and/or Wpromoters on the structures of mixed MoVW oxides and their catalytic properties are investigated by XRD, TEM, SEM, in situ Raman spectroscopy and TPRS. V addition causes high propene conversions and the formation of an oxide with Mo5O14 structure is observed. Minor amounts of W in the molybdenum oxides matrix inhibit structural reorganisation processes which is explained by preferred octahedral coordination of the redox stable W. The combined XRD, TEM and Raman spectroscopic identification of the oxide of Mo5O14 structure points to its relevant role for the selective propene oxidation. A resonance Raman effect is proven to be operative in oxygen defective molybdenum oxides. For an excitation wavelength of 632 nm (1.96 eV), the observed Raman cross section varies as a function of the degree of reduction of five different MoO3 x samples. A model of the electronic transitions in MoO3 x based on crystal field theory explains the electronic transitions observed by DR-UV/VIS spectroscopy. The observed resonant Raman scattering is coupled to the IVCT transition at about 2 eV arising from oxygen vacancies present in the materials. Due to the local nature of the absorption process, the developed model is valid for intermediate oxides too. Resonant Raman scattering was proven for Mo4O11 and MoO2 too. Hence, the experimentally observed Raman intensity bears information about the degree of reduction of the molybdenum oxide. In situ Raman spectroscopy of MoO3 x catalysts during propene partial oxidation indicates that the propene conversions and the selectivities are a function of the degree of reduction of the catalyst. It is attempted to control the spatial elemental distribution within the polyoxometalate catalyst precursor by the formation of molecularly defined species in solution by addition of acetate, oxalate or tartrate. Due to the low stability of the acetate complexes, the addition of acetate only leads to minor amounts of monomeric species beside mixed and pure polyoxometalates. The formation of stable monomeric oxometalate oxalate and tartrate complexes is observed, which are expected to lead to the formation of a catalyst precursor with a homogeneous elemental distribution.

The late Cambrian–Early Ordovician siliciclastic sequence in central northern Tasmania represents an excellent case study to document the transition from an active rifting succession to a post-rift system. Initial rifting created a complex system of half-graben, providing accommodation space for large volumes of basement-derived material. These half-graben were initially filled with coarse-grained alluvial fan sediments and a thick, syn-rift, low-sinuosity, multiple-channel braided fluvial succession. The uppermost sequence comprises a post-rift, regional, shallow marine, transgressive sandstone and mudstone package. <br> <br> Six broad lithofacies have been recognised in the late Cambrian–Early Ordovician siliciclastic sequence: Four lithofacies are interpreted to represent terrestrial alluvial fan (displaying debris flow, sheet-flow and channel-flow geometries) and braided fluvial deposits (transitioning between proximal and distal sedimentological characteristics upsection), while two represent marginal-marine (tidal) to shallow marine environments (dominated by heterolithic bedding, and varying degrees and styles of bioturbation). Adjacent lithofacies often display marked lateral variations in both thickness and grain-size, suggesting accelerated changes in the volume of sediment flux and/or the rate of tectonic subsidence, particularly of the basal, coarse-grained, conglomerate-rich sequences where the varying composition and texture suggests considerable relief on hinterland palaeotopography. <br> <br> A revised stratigraphic framework for the late Cambrian–Early Ordovician siliciclastics of central northern Tasmania is presented based on lithofacies and lithofacies associations. The Roland Formation forms the lower stratigraphic unit and comprises generally fining-up, conglomerate-dominated alluvial fan and proximal, low-sinuosity, multiple-channel braided river successions associated with syn-rift sedimentation. A hiatus in the extensional tectonics is recorded by a regionally extensive unconformity between the Roland Formation and the overlying Moina Formation, named the intra-Owen Group Unconformity. The Moina Formation marks a transition from high energy, terrestrial, conglomerate-rich sequences of the Roland Formation, to more moderate energy sandstone-dominated sequences that typically display an increasing marine signature towards the top of the unit. Three distinct sedimentary successions are recognised within the Moina Formation, and member status has been given to each of these since the distribution and relationships between these stratal packages is fundamental in understanding the mature phase of the late Cambrian–Early Ordovician extensional event, and the post-rift transgressive system. The terrestrial braided fluvial, channelised sandstone and minor conglomerate of the Badgers Range Member are superseded by interbedded, fine- to medium-grained sandstone deposited in a tidal regime (Deloraine Member) that is in turn overlain by thinly bedded, heavily bioturbated, shallow marine mudstone and fine-grained sandstone (Caroline Creek Member). <br> <br> Sedimentological principles, including provenance studies and palaeocurrent analysis, are used to document the basin configuration and structural framework of the rift system. The spatial and temporal migration of depocentres can be documented such as in the Badgers Range where two fining-up successions are separated by the intra-Owen Group Unconformity with the lower quartzite-dominated Roland Formation being succeeded by the chert lithic-rich Moina Formation. It is apparent that depocentres were compartmentalised by topographic highs, and the sedimentary fill of these depocentres reflects the lithological composition of these highs. In addition, the observed general decrease in grain-size is not uniform, and there are numerous changes in sedimentation as a consequence of fluxes in uplift and subsidence in the source and basinal areas. <br> <br> The distribution of thickly-bedded, coarse-grained conglomerate sequences, the juxtaposition of differing lithological successions, and the construction of geological cross-sections give insights to the location and position of several major bounding faults. These faults were subsequently reactivated during the Early–Middle Devonian Tabberabberan Orogeny, and a significant amount of reverse movement is recorded on the basis of structural restorations. <br> <br> Regional palaeogeographic reconstructions of the Early Ordovician (Tremadocian; 480 Ma) (Cocks, 2001) demonstrate that Tasmania was located outboard of the eastern margin of Gondwana, situated north of the palaeoequator between latitudes 10° and 20°. The late Cambrian through Ordovician tropical climate was dominated by the influence of a strong, extended greenhouse effect. The lack of sediment stabilisation by plants and rootlets in this Early Palaeozoic, vegetation-free, terrestrial landscape resulted in alluvial and fluvial depositional processes being markedly different from their present day counterparts, since vegetation exerts significant influence on a variety of environmental factors. In particular, energy levels and run-off rates would have been greater, fluctuations in energy levels more extreme, and high energy events more frequent. These conditions would likely have been amplified by the relatively high precipitation rates associated with the tropical palaeogeographic position of Tasmania. Likewise, wind processes would have played a far greater role than present with regards to the removal of clay- and silt-sized particles from the depositional setting, and this may explain the paucity of claystone and siltstone in these terrestrial environments.

Bulk Metallic Glasses (BMGs) have been drawing increasing attention in recent years due to their scientific and engineering significance. A great deal of effort in this area has been devoted to developing BMGs in different alloy systems. BMGs based on certain late transition metals (e.g. Fe, Co, Ni, Cu) have many potential advantages over those based on early transition metals. These include even higher strength and elastic modulii, and lower materials cost, to name a few, which are highly preferable for a broad application of BMGs as engineering materials. Nevertheless, these ordinary-late-transition-metal-based BMGs generally have quite limited glass-forming ability (GFA). In particular, for the Ni-based and Cu-based alloys reported prior to this research, the maximum casting thickness allowed to retain their amorphous structures is only ~2 mm (or lower) and ~5 mm (or lower), respectively. During this research it was first found that certain quinary Ni-based alloys in the Ni-Cu-Ti-Zr-Al system can be cast into 5 mm diameter amorphous rods. This critical casting thickness is the highest for any reported Ni-based BMG’s indicating that these alloys are the easiest metallic glass formers based on Ni discovered to date. Secondly, certain binary alloys in the Cu-Zr and Cu-Hf systems were found to form bulk glasses with casting thickness as high as 2 mm. The discovery of these binary BMGs was very surprising since it had been widely considered that only multi-component (containing at least three elements) alloys could form bulk metallic glasses. These new binary BMGs provide interesting subjects for future theoretical studies such as molecular dynamics simulations since they possess both the simplicity of binary alloys and the good GFA of multi-component BMGs. In fact, these binary BMGs have led to the third but perhaps the most significant progress during this research, i.e., the discovery of a family of Cu-based BMGs in the Cu-Zr-Al-Y system which possess a critical casting thickness up to 1 cm. These quaternary Cu-based alloys, together with some complicated Fe-based alloys reported by two other groups during the course of this research, are the first centimeter level BMGs based on the ordinary late transition metals. This thesis first reviews the fundamentals related to BMG development, then reports in detail the formation and properties of the above-mentioned binary and multi-component BMGs based on Ni and Cu. A generalized geometric model for the critical-value problem of nucleation developed in this research is also presented.

In the field of computer vision and image processing CBIR (Content based image retrieval) system become an active research area. The CBIR technique involves feature extraction based on image features (like color, shape and texture) and defining rules for comparing image. A classifier is also used by authors for image classification. For this purpose an image database taken from James Z. Wang research group website is used by many authors. Due to enormous increase in the size of database traditional methods of image retrieval by using keywords does not provide accurate result. Thus the concept of CBIR system is emerged. In last couple of years many method of feature extraction is developed but since now researcher are not able to develop a system which has gained high accuracy of human visual perception in distinguishing pictures. Therefore earlier researchers have concentrated to increase the retrieval accuracy of CBIR system by using different combination of features. The objective of this paper is to present review on a CBIR technique which combined four kinds of features set including every kind of feature for developing feature vector. Color based features (histogram), statistical features (mean, median and standard deviation), spatial features (Gabor feature and wavelet features) and texture features (GLCM) is used for forming feature vector. A SVM classifier is used with CBIR system to increase retrieval accuracy. SVM (support vector machine) is a machine learning classifier which provides good classification accuracy. Keywords: CBIR, SVM, NN, GLCM, HSV Color space, Euclidean distance matrix.

BackgroundEarly identification of Alzheimer’s disease or mild cognitive impairment can help guide direct prevention and supportive treatments, improve outcomes, and reduce medical costs. Existing advanced diagnostic tools are mostly based on neuroimaging and suffer from certain problems in cost, reliability, repeatability, accessibility, ease of use, and clinical integration. To address these problems, we developed, evaluated, and implemented an early diagnostic tool using machine learning and non-imaging factors.Methods and resultsA total of 654 participants aged 65 or older from the Nursing Home in Hangzhou, China were identified. Information collected from these patients includes dementia status and 70 demographic, cognitive, socioeconomic, and clinical features. Logistic regression, support vector machine (SVM), neural network, random forest, extreme gradient boosting (XGBoost), least absolute shrinkage and selection operator (LASSO), and best subset models were trained, tuned, and internally validated using a novel double cross validation algorithm and multiple evaluation metrics. The trained models were also compared and externally validated using a separate dataset with 1,100 participants from four communities in Zhejiang Province, China. The model with the best performance was then identified and implemented online with a friendly user interface. For the nursing dataset, the top three models are the neural network (AUROC = 0.9435), XGBoost (AUROC = 0.9398), and SVM with the polynomial kernel (AUROC = 0.9213). With the community dataset, the best three models are the random forest (AUROC = 0.9259), SVM with linear kernel (AUROC = 0.9282), and SVM with polynomial kernel (AUROC = 0.9213). The F1 scores and area under the precision-recall curve showed that the SVMs, neural network, and random forest were robust on the unbalanced community dataset. Overall the SVM with the polynomial kernel was found to be the best model. The LASSO and best subset models identified 17 features most relevant to dementia prediction, mostly from cognitive test results and socioeconomic characteristics.ConclusionOur non-imaging-based diagnostic tool can effectively predict dementia outcomes. The tool can be conveniently incorporated into clinical practice. Its online implementation allows zero barriers to its use, which enhances the disease’s diagnosis, improves the quality of care, and reduces costs.